Search

Studies of dairy cow longevity usually focus on the animal life after first calving, with few studies considering early life conditions and their effects on longevity. The objective was to evaluate the effect of birth conditions routinely collected by Dairy Herd Improvement agencies on offspring longevity measured as length of life and length of productive life. Lactanet provided 712,890 records on offspring born in 5,425 Quebec dairy herds between January 1999 and November 2015 for length of life, and 506,066 records on offspring born in 5,089 Quebec dairy herds between January 1999 and December 2013 for length of productive life. Offspring birth conditions used in this study were calving ease (unassisted, pull, surgery, or malpresentation), calf size (small, medium, or large), and twinning (yes or no). Observations were considered censored if the culling reason was "exported," "sold for dairy production," or "rented out" as well as if the animals were not yet culled at the time of data extraction. If offspring were not yet culled when the data were extracted, the last test-day date was considered the censoring date. Conditional inference survival trees were used in this study to analyze the effect of offspring birth conditions on offspring longevity. The hazard ratio of culling between the groups of offspring identified by the survival trees was estimated using a Cox proportional hazard model with herd-year-season as a frailty term. Five offspring groups were identified with different length of life based on their birth condition. Offspring with the highest length of life [median = 3.61 year; median absolute deviation (MAD) = 1.86] were those classified as large or medium birth size and were also the result of an unassisted calving. Small offspring as a result of a twin birth had the lowest length of life (median = 2.20 year; MAD = 1.69) and were 1.52 times more likely to be culled early in life. Six groups were identified with different length of productive life. Offspring that resulted from an unassisted or surgery calving and classified as large or medium when they were born were in the group with the highest length of productive life (median = 2.03 year; MAD = 1.63). Offspring resulting from a malpresentation or pull in a twin birth were in the group with the lowest length of productive life (median = 1.15 year; MAD = 1.11) and were 1.70 times more likely to be culled early in life. In conclusion, birth conditions of calving ease, calf size, and twinning greatly affected offspring longevity, and such information could be used for early selection of replacement candidates.

In the original article [...]

Simple Summary This paper discusses mechanisms to promote the sustained adoption of integrated decision support systems (IDSS) in dairy farming. Integrated DSS are those that receive continuous data streams from on-farm and off-farm data collection systems and integrate them to produce insightful and effective outcomes. Multiple limiting factors are preventing the use and adoption of existing IDSS and could also prevent the adoption of upcoming IDSS. In summary, these factors include value proposition, practical application, ease of use, and data governance issues related to the IDSS. Therefore, awareness and further industry-wide discussion of these components could help in the development and sustained adoption of IDSS in order to facilitate the complex process of decision making. As a commentary paper, the emphasis is on the authors’ extensive experience, expertise, and practical knowledge. Abstract Dairy farm decision support systems (DSS) are tools which help dairy farmers to solve complex problems by improving the decision-making processes. In this paper, we are interested in newer generation, integrated DSS (IDSS), which additionally and concurrently: (1) receive continuous data feed from on-farm and off-farm data collection systems and (2) integrate more than one data stream to produce insightful outcomes. The scientific community and the allied dairy community have not been successful in developing, disseminating, and promoting a sustained adoption of IDSS. Thus, this paper identifies barriers to adoption as well as factors that would promote the sustained adoption of IDSS. The main barriers to adoption discussed include perceived lack of a good value proposition, complexities of practical application, and ease of use; and IDSS challenges related to data collection, data standards, data integration, and data shareability. Success in the sustainable adoption of IDSS depends on solving these problems and also addressing intrinsic issues related to the development, maintenance, and functioning of IDSS. There is a need for coordinated action by all the main stakeholders in the dairy sector to realize the potential benefits of IDSS, including all important players in the dairy industry production and distribution chain.

Simple Summary The ability of farms to produce milk sustainably is closely related to dairy cow longevity, i.e., the length of productive life. However, longevity is a very complex feature that depends on all the aspects of the lifespan of a cow and there is no standard definition nor metric to measure it. Measuring longevity is important because it influences the profitability and the environmental impact of farms as well as the welfare of the animals. The objectives of this paper were to review metrics used to measure longevity and describe its status among high milk-producing countries. Increasing dairy cow longevity would imply that an animal has an early age at first calving and a long and profitable productive life. Combining age at first calving, length of productive life, and margin over all (available) costs provides a complete evaluation of longevity. This paper also shows that dairy cow longevity has decreased in most high milk-producing countries over time, which confirm the concerns voiced by the dairy industry and other stakeholders. Increasing cow longevity would reduce health costs and increase cow profitability while improving both animal welfare and quality of life, contributing to a more sustainable dairy industry. Abstract The ability of dairy farmers to keep their cows for longer could positively enhance the economic performance of the farms, reduce the environmental footprint of the milk industry, and overall help in justifying a sustainable use of animals for food production. However, there is little published on the current status of cow longevity and we hypothesized that a reason may be a lack of standardization and an over narrow focus of the longevity measure itself. The objectives of this critical literature review were: (1) to review metrics used to measure dairy cow longevity; (2) to describe the status of longevity in high milk-producing countries. Current metrics are limited to either the length of time the animal remains in the herd or if it is alive at a given time. To overcome such a limitation, dairy cow longevity should be defined as an animal having an early age at first calving and a long productive life spent in profitable milk production. Combining age at first calving, length of productive life, and margin over all costs would provide a more comprehensive evaluation of longevity by covering both early life conditions and the length of time the animal remains in the herd once it starts to contribute to the farm revenues, as well as the overall animal health and quality of life. This review confirms that dairy cow longevity has decreased in most high milk-producing countries over time and its relationship with milk yield is not straight forward. Increasing cow longevity by reducing involuntary culling would cut health costs, increase cow lifetime profitability, improve animal welfare, and could contribute towards a more sustainable dairy industry while optimizing dairy farmers’ efficiency in the overall use of resources available.

The series of events leading to the decision to cull a cow is complex, involving both individual-level and herd-level factors. While the decision is guided by financial returns, it is also influenced by social and psychological factors. Research studies on the motivational and behavioural aspects of farmers' decision utility are sparse, and nonexistent regarding culling expectations and its decision process. Our goal was to identify shared criteria on culling decisions held by dairy producers and farm advisers, with the help of the Q-methodology. Forty-one dairy producers and 42 advisers (17 veterinarians, 13 feed mill advisers, and 12 dairy herd improvement (DHI) advisers) undertook a Q-sort with 40 statements that represented a range of views about cow and herd health, production performance, management issues, and material factors that might impact their culling decision-making process. The sorts were analysed by-person using factor analysis and oblimin rotation. A single view on culling could be identified among dairy producers that can be extended to dairy farm advisers, who showed two variations of the same well-structured, uni-dimensional decision-making process. Udder health, milk production performance, and milk quota management were the key criteria for the culling decision. Farm management parameters (debts, amortization, employees, milking parlour capacity, herd size) did not play any role in the decision process. Three key differences were, however, identified between producers and the two types of advisers. One group of advisers followed the recommendations from mathematical models, where pregnancy is a major determinant of a cow's value. They assessed the cow in a more abstract way than did the other participants, still taking into account udder health and milk production, but adding economic considerations, like the availability of financial incentives and an evaluation of the post-partum health of the cow. Dairy producers were also more concerned about producing healthy and safe milk, which might reflect a different value given to dairy farming than by advisers. Very different degrees of importance were given to animal welfare by the three groups, which could represent different views on the attributed relationships between dairy farmers and their animals. Our findings suggest that dairy producers and their advisers hold a general common view regarding culling decision-making. However there are significant differences between producers and advisers, and among advisers. Understanding and managing these differences is important for assisting the change management processes required to increase farm profitability, and call for further investigation.

The objective of this study was to examine the effect of feeding systems [component and total mixed rations (TMR)] and dietary grain sources (barley, commercial concentrate, corn grain, and high-moisture corn) on lactation characteristics and milk composition. A total of 852,242 test-day records, information on animal characteristics, feed composition, and feeding systems from 104,129 Holstein cows in 4,319 herds covering a period of 5 yr were obtained from Quebec's Dairy Herd Improvement Association (Valacta). We performed descriptive statistics and graphical representations of the data for each type of feeding system and grain source by parity (1 to 3). The milk records were binned in 15-d in milk blocks. Mixed models using a combination of forward and backward stepwise selections were developed to predict milk and milk component yields. The TMR-fed cows had greater yield of milk, fat, protein, and lactose and lower milk urea N (MUN) concentration than component-fed cows at all parities. Cows fed a TMR had higher peak milk yields and greater persistency after peak lactation compared with component-fed cows. In addition, greater yields of milk fat and protein from peak to mid-lactation were found in TMR- versus component-fed cows. In general, greater milk fat and protein yields as well as lower MUN concentration were observed in cows fed corn grain or high-moisture corn compared with barley or commercial concentrate, but parity influenced these relationships. The feeding system by day in milk blocks interaction was significant in models of milk and components yields for all parities, but only for second-lactation cows for MUN concentration. This means that effect of TMR and component feeding differs with stage of lactation. In conclusion, feeding TMR and corn-based diets are associated with greater yield of milk and milk components under commercial conditions.

A retrospective cohort study was conducted using production and health records from Quebec dairy herds which were integrated with the objectives of calculating different lifetime profitability measures and developing analyses of the impact of selected reproduction and health variables on profitability. The dataset for the study consisted of lifetime records for a total of 13 668 Holstein cows from 113 herds and in cohorts of cows calving for the first time between 2000 and 2009. To calculate profit, all the revenues and costs were accumulated on a lifetime basis, and cumulative lifetime profit and cumulative lifetime profit adjusted for the opportunity cost of postponed replacement were selected for analysis. There were statistically significant effects in both profit measures because of age at first calving, cumulative number of days in milk, and cumulative number of days dry as well as the number of health events accumulated by animal lifetime (clinical mastitis, reproductive problems, and feet and legs...

Several health disorders, such as milk fever, displaced abomasum, and mastitis, as well as impaired reproductive performance, are known risk factors for the removal of affected cows from a dairy herd. While cow-level risk factors are well documented in the literature, herd-level associations have been less frequently investigated. The objective of this study was to investigate the effect of cow- and herd-level determinants on variations in culling risk in Quebec dairy herds: whether herd influences a cow's culling risk. For this, we assessed the influence of herd membership on cow culling risk according to displaced abomasum, milk fever, and retained placenta. A retrospective longitudinal study was conducted on data from dairy herds in the Province of Quebec, Canada, by extracting health information events from the dairy herd health management software used by most Quebec dairy producers and their veterinarians. Data were extracted for all lactations starting between January 1st and December 31st, 2010. Using multilevel logistic regression, we analysed a total of 10,529 cows from 201 herds that met the inclusion criteria. Milk fever and displaced abomasum were demonstrated to increase the cow culling risk. A minor general herd effect was found for the culling risk (i.e. an intra-class correlation of 1.0% and median odds ratio [MOR] of 1.20). The proportion of first lactation cows was responsible for this significant, but weak herd effect on individual cow culling risk, after taking into account the cow-level factors. On the other hand, the herd's average milk production was a protective factor. The planning and management of forthcoming replacement animals has to be taken into consideration when assessing cow culling risks and herd culling rates.

Health disorders, such as milk fever, displaced abomasum, or retained placenta, as well as poor reproductive performance, are known risk factors for culling in dairy cows. Clinical mastitis (CM) is one of the most influential culling risk factors. However the culling decision could be based either on the disease status or on the current milk yield, milk production being a significant confounder when modelling dairy cow culling risk. But milk yield (and somatic cell count) are time-varying confounders, which are also affected by prior CM and therefore lie on the causal pathway between the exposure of interest, CM, and the outcome, culling. Including these time-varying confounders could result in biased estimates. A marginal structural model (MSM) is a statistical technique allowing estimation of the causal effect of a time-varying exposure in the presence of time-varying covariates without conditioning on these covariates. The objective of this paper is to estimate the causal effect on culling of CM occurring between calving and 120 days in milk, using MSM to control for such time-varying confounders affected by previous exposure. A retrospective longitudinal study was conducted on data from dairy herds in the Province of Quebec, Canada, by extracting health information events from the dairy herd health management software used by most Quebec dairy producers and their veterinarians. The data were extracted for all lactations starting between January 1st and December 31st, 2010. A total of 3952 heifers and 8724 cows from 261 herds met the inclusion criteria and were used in the analysis. The estimated CM causal hazard ratios were 1.96 [1.57-2.44] and 1.47 [1.28-1.69] for heifers and cows, respectively, and as long as causal assumptions hold. Our findings confirm that CM was a risk factor for culling, but with a reduced effect compared to previous studies, which did not properly control for the presence of time-dependent confounders such as milk yield and somatic cell count. Cows experienced a lower risk for CM, with milk production having more influence on culling risk in cows than heifers.

The relationship between cows' health, reproductive performance or disorders and their longevity is well demonstrated in the literature. However these associations at the cow level might not hold true at the herd level, and herd-level variables can modify cow-level outcomes independently of the cows' characteristics. The interaction between cow-level and herd-level variables is a relevant issue for understanding the culling of dairy cows. However it requires the appropriate group-level variables to assess any contextual effect. Based on 10 years of health and production data, the objectives of this paper are:(a) to quantify the culling rates of dairy herds in Quebec; (b) to determine the profiles of the herds based on herd-level factors, such as demographics, reproduction, production and health indicators, and whether these profiles can be related to herd culling rates for use as potential contextual variables in multilevel modelling of culling risk. A retrospective longitudinal study was conducted on data from dairy herds in Quebec, Canada, by extracting health information events from the dairy herd health management software used by most Quebec producers and their veterinarians. Data were extracted for all lactations taking place between January 1st, 2001 and December 31st, 2010. A total of 432,733 lactations from 156,409 cows out of 763 herds were available for analysis. Thirty cow-level variables were aggregated for each herd and years of follow-up, and their relationship was investigated by Multiple Factor Analysis (MFA). The overall annual culling rate was 32%, with a 95% confidence interval (CI) of [31.6%,32.5%]. The dairy sale rate by 60 days in milk (DIM) was 3.2% [2.8%,3.6%]. The annual culling rate within 60 DIM was 8.2% [7.9%,8.4%]. The explained variance for each axis from the MFA was very low: 14.8% for the first axis and 13.1% for the second. From the MFA results, we conclude there is no relationship between the groups of herd-level indicators, demonstrating the heterogeneity among herds for their demographics, reproduction and production performance, and health status. However, based on Principal Component Analysis (PCA), the profiles of herds could be determined according to specific, single, herd-level indicators independently. The relationships between culling rates and specific herd-level variables within factors were limited to livestock sales, proportion of first lactation cows, herd size, proportion of calvings occurring in the fall, longer calving intervals and reduced 21-day pregnancy rates, increased days to first service, average age at first calving, and reduced milk fever incidence. The indicators found could be considered as contextual variables in multilevel model-building strategies to investigate cow culling risk.

There is growing interest among the public in farm animal welfare and a need for methods to assess animal welfare on farm. A survey on calf rearing practices that might affect dairy calf welfare was performed via a 1-h interview on 115 dairy farms (mean +/- SD: herd size=52.5+/-20.9 cows; milk production=8,697+/-1,153L) distributed throughout the province of Quebec. Despite frequent recommendations, many dairy producers continue to use management practices that increase the health risks of milk-fed calves. Major risk factors for poor calf welfare identified were 1) no use of calving pen in 51.3% of herds and low level of surveillance of calvings, especially at nighttime (once every 12h), 2) no disinfection of newborn's navel in 36.8% of herds, and delayed identification and, hence, calf monitoring (3 d), 3) 15.6% of farms relied on the dam to provide colostrum and none checked colostrum quality or passive transfer of immunity, 4) dehorning and removal of extra teats proceeded at late ages (6.4 wk and 6.7 mo, respectively) and without adequate pain control, 5) use of traditional restrictive milk feeding and waste milk distributed to unweaned calves without precaution in 48.2% of herds, 6) abrupt weaning performed in 16.5% of herds, and 7) calves housed individually in 87.9% of herds, and most inappropriate housing systems (crate=27.0%, tie-stall=13.9%, attached against a wall=5.7%) remained. This risk factor assessment was the first step in an intervention strategy to improve calf welfare on dairy farms.

The reduction in milk production caused by subclinical mastitis in dairy cattle was assessed through the regression of test day milk yield on log-transformed somatic cell counts (LnSCC). Data was obtained from Valacta, Quebec, and a total of 312,756 test day records from Ayrshire cows and 1,869,785 test day records from Holstein cows were included in the analyses. A segmented regression was fitted to estimate the cutoff point in the LnSCC scale where milk yield starts to be affected by mastitis. The statistical model used to explain daily milk yield included the effects of herd–year-season of test (random), days in milk, age at calving and LnSCC, and analyses were performed by breed, parity and stage of lactation. The cutoff point where milk yield starts to be affected by changes in LnSCC was estimated from data to be around 2 (approximately 7400 cells/mL) for Canadian Ayrshires and Holsteins. Milk losses per unit increase in LnSCC varied from 0.55 to 0.84 kg/day in first lactation Ayrshires, from 0.33 to 0.55 kg/day in first lactation Holsteins, from 0.74 to 2.45 kg/day in adult Ayrshires and from 0.77 to 1.78 kg/day in adult Holsteins. Daily milk losses caused by changes in LnSCC were dependent on breed, parity and stage of lactation, and these factors should be considered when estimating losses associated with subclinical mastitis.

As a growing world population and diminishing physical resources continue to count among the biggest challenges facing food supply and security, the provision of quality animal-food products is examined in terms of available technologies to increase their supply in an equitable manner. Improvements through genetic selection and modification, reproductive technologies, and a better understanding of the nutritional needs for various animal production systems are presented. The need for sustainable systems is also discussed within a framework of ethical animal treatment and welfare. These advances are also considered in the context of economic, health, and environmental considerations (where goals often seem to be in conflict), and from the standpoint of developed- and developing-world attitudes to the production and consumption of animal food products. It is proposed that scientific advances need to be better explained to societies that are becoming more critical of food choices, their benefits, and the systems under which they are produced.

The objective was to provide an overview of heifer growth in Quebec and analyze its relationship with age at first calving. Heifer body weights, wither heights, body condition scores, and breeding events collected from 1993 to 2003 were obtained and combined with first lactation data, mature body weights, type classification final score and frame/capacity, and estimated breeding values for milk, conformation, frame/capacity, and Lifetime Profit Index. Analyses were restricted to 39 animal-level variables. The resulting data set consisted of 44 989 Holsteins and 2294 Ayrshires with data available for at least one variable. The average age at first calving was 26.5 and 27.1 mo for Holsteins and Ayrshires, respectively. For Holsteins, average daily gains in body weights were 0.79, 0.89, 0.87, and 0.70 kg d-1 for 0 to 4 mo (birth to weaning), 4 to 10 mo (to target onset of puberty), 10 to 14 mo (to first breeding), and 14 mo to pre-calving periods of growth, respectively. Body weights at 14 mo, conception, and post calving were 398, 464, and 601 kg, respectively, and relative to mature body weight were 59, 68 and 89%, respectively. Correlations with age at first calving, corrected for herd and year of birth effects, were positive for body weight at conception (0.69), body weight at first calving (0.31), and number of breedings (0.37), and negative for the average daily gain from conception to post calving (-0.38) and body weight at 14 mo relative to mature body weight (-0.20). For Ayrshires, body weights and average daily gains were lower than for Holsteins, but body weights relative to mature body weight and correlations were generally similar. Holsteins and Ayrshires calved substantially earlier and at a heavier body weight compared with Quebec data from the 1980s. The results of this study suggest that heifer growth may not be a limiting factor in reducing the average age at first calving in Québec. Key words: Dairy cattle, field-recorded data, heifer, growth

The objective of this study was to evaluate the potential usage, benefits, and limitations of using 3D graphics in an interactive exploration of multi-dimensional dairy data. A software prototype was, therefore, designed, consisting of two modules: a histogram to view joint frequency distributions for two dimensions of the data set at a time; and a scatter plot to explore clouds of data points plotted along three dimensions. A number of interaction techniques were implemented including chart zooming and translation, brushing of chart objects to display details, rapid attribute selection per chart coordinate, and dynamic data filtering. Chart rotation was implemented in order to analyze the 3D representations optimally. The feasibility of developing such prototypes was demonstrated in terms of data access, filtering, flexibility, and functionality under standard personal computer resources. However, performance tests with the 3D scatter plot revealed that in order for smooth interactions to occur, data needed to be limited to a few thousand records. The two resulting modules were seen as potentially useful in the support of various on-farm decision-making activities including the monitoring of cow performance for the most recent milk test, and the benchmarking of herd performance, based on annual economics and milk recording data.

The objectives of this research were: (a) to perform a survey of current research in the area of multi-agent systems in order to learn more about how they could be designed and implemented; (b) to investigate the feasibility of such an approach for agriculture, based on an integration of currently existing technologies; and, more specifically (c) to assess the potential of a multi-agent approach in the context of decision support for dairy production. The results of this work highlighted a number of key concepts in multi-agent system design, including the importance of selecting an appropriate system architecture for agent coordination (e.g., peer-to-peer, federated, or blackboard-based) and the need for well-defined agent communication methods (language and ontology). Alternative technologies, used in the implementation of multi-agent systems (e.g., communication protocols and distributed computing methods), were also explored. Lastly, a case study was carried out, in which some of the discussed technologies were tested and implemented to create a multi-agent heifer management system. The system consists of two different types of agents and several databases, and was implemented on a PC-based network. The agents work together to synthesize data about heifer development from different sources and to present this to the user in a graphical format. The system demonstrates the feasibility of applying an agent-based approach, using currently available technology, to problems such as dairy herd management in which a distributed decision-support solution is often required. It is concluded that the constraints for the implementation of multi-agent systems do not appear to be of a technological nature; the challenge seems to be more one of defining and accepting a common ontology and communication language by members of a given industry. In addition, large-scale distributed systems will require sophisticated agent-coordination methods to ensure robust and efficient operation.

Machine-learning techniques are increasingly used to deal with a variety of problems in agriculture. However, challenges with the application of machine-learning, such as analyzing the performance achieved through learning from small data sets, still remain. This study focused on using graphical and statistical techniques to analyze the results of machine-learning experiments involving data preprocessing and algorithm tuning. Data consisted of 1428 cases that were classified by a dairy-nutrition specialist as outliers (34 cases) or non-outliers. The performance of classifiers, generated with decision-tree induction, was estimated through ten-fold cross validation. Relative operating characteristic (ROC) curves were used to visualize the achieved trade-offs between correctly classifying positive and negative cases. A performance index, representing the mean true positive rate of these curves for a limited range of false positive rate values, was developed to facilitate comparison among classification schemes. Analysis of variance (ANOVA) was used to determine whether real differences existed for the expected performance on new data among the different combinations of data preprocessing and algorithm configurations evaluated in this study. In terms of data preprocessing, randomly assigning herds to the folds of the cross validation did not perform significantly differently from assigning cases to folds, while using a special value to indicate irrelevant attribute values significantly improved the performance over treating these values as unknown. Tuning the configuration of the decision-tree induction algorithm significantly improved the classification performance. The application of ten-fold cross validation in combination with ROC curves and ANOVA was found to be useful in analyzing the results of machine-learning experiments involving decision-tree induction and small data sets. These methods could also be used with other machine-learning techniques such as artificial neural networks and instance-based learning.

Machine learning has been identified as a promising approach to knowledge-based system development. This study focused on the use of decision-tree induction for knowledge acquisition to filter individual-cow lactations for group-average lactation curve analysis. Data consisted of 1428 cases, classified by a dairy-nutrition specialist as outliers (34 cases) or non-outliers. The classification performance was estimated through 10-fold cross validation. A relative operating characteristic curve was used to visualize the achievable range of trade-offs between correctly classifying positive and negative cases. A series of three final decision trees with increasing tendency of classifying a lactation as outlier was induced from the entire data set. For these trees, the expected true positive rates were 52, 68 and 92%, at false positive rates of 1.5, 3.5 and 8.6%, respectively. However, due to the low prevalence of outlier lactations (cases), this performance was associated with many false positives. The performance of individual decision nodes was tested against the entire data set to identify potentially counter-intuitive nodes resulting from overspecialization to the training data. The specialist reviewed the final trees and adjusted two decision nodes. This study suggests that although the input from a domain specialist remains important, decision-tree induction is a useful technique to support knowledge acquisition involved in the removal of outlier lactations.

A case-acquisition and decision-support system was developed to support the analysis of group-average lactation curves and to acquire example cases from domain specialists. This software was developed through several iterations of a three-step approach involving 1) problem analysis and formulation in consultation with two dairy nutrition specialists; 2) development of a case-acquisition and decision-support prototype by the system developer; and 3) use of the prototype by the domain specialists to analyze and classify milk-recording data from example herds. The overall problem was decomposed into three subproblems: removal of outlier tests and lactation curves of individual cows; interpretation of group-average lactation curves; and diagnosis of detected abnormalities at the herd level through the identification of potential management deficiencies. For each subproblem, a software module was developed allowing the user to analyze both graphical and numerical performance representations and classify these representations using predefined linguistic descriptors. The example-based method for the development of the program proved to be very useful, facilitating the communication between system developer and domain specialists, and allowing the specialists to explore the appropriateness of the various prototypes developed. The resulting software represents a formalization of the approach to group-average lactation curve analysis, elicited from the two domain specialists. In future research, the case-acquisition and decision-support system will be complemented with knowledge to automate identified classification tasks, which will be captured through the application of machine-learning techniques to example cases, acquired from domain specialists using the software.

In fuzzy logic, the calibration of membership functions is often cumbersome and the use of univariate models, such as triangular or trapezoidal functions, is not always ideal. In this study, artificial neural networks were used as multivariate models for the fuzzification of milk yield, with the aim of providing a more convenient approach than traditional techniques. The objective was to develop a neuro-fuzzifier that would mimic an expert’s process of assigning dairy production records to fuzzy milk-yield sets and assessing their corresponding degrees of membership. Data consisted of 313 dairy production records fuzzified by an expert, according to herd-average 305-day milk yield, cow lactation number, days in lactation, standard milk yield on test day, test day milk-yield and deviation of test day production from a standard value. Five fuzzy sets for milk yield were predetermined as being very low, low, medium, high and very high. Two neuro-classifiers were first trained to identify the sets to which a record belonged and then, four specialized networks were devised to predict the corresponding degrees of membership. In order to evaluate this approach, network classification and predictions were compared with those obtained from a previously developed univariate model and results showed that it was generally better in both the classification of milk yields into fuzzy sets, and the determination of corresponding degrees of membership. These results suggest that the approach of neuro-fuzzification may be superior to traditional methods for applications that require multivariate fuzzifiers.

A data set comprising milk-recording and conformation data was used to investigate the usefulness of artificial neural networks in detecting influential variables in the prediction of incidences of clinical mastitis. Specifically, these data contained test-day records from dairy herd analysis, phenotypic cow scores for conformation and genetic conformation proofs for cows and their sires. The data were analysed using the milk-recording data only, the conformation data only, and a combination of the two. Results from sensitivity analyses, performed with trained neural nets, indicated that stage of lactation, milk yield on test day, cumulative milk yield and somatic cell count were the major production factors influencing the ability to detect the occurrence of clinical mastitis. Among the conformation traits, such variables as phenotypic scores for rear-teat placement, dairy character and size, cow proof for dairy character, sire reliability for final score and sire proofs for pin-setting (desirability) and loin strength were found to have some influence on the network's predictive ability, although they were all very minor in relation to the production variables mentioned. As a group, cow genetic proofs seemed more important than either sire genetic proofs or cow phenotypic scores. Given the neural network's general abilities to determine the major factors related to the presence or absence of mastitis on a given test day, it may be appropriate to investigate the possibility of using this technology for actual prediction purposes. Key words: Artificial neural networks, clinical mastitis, milk-recording data, conformation traits, sensitivity analysis, milk production

A back-propagation artificial neural network was employed to detect clinical mastitis using a file of 460,474 test day records. Two data files were created to train the artificial neural networks, containing a relatively large (1:1) ratio and a relatively small (1:10) ratio in the incidence to non-incidence of clinical mastitis. These ratios were applied to each of two input file designs; one comprised variables that are traditional in the modeling of mastitis (e.g., age, stage of lactation and somatic cell count) and a second included additional variables (e.g., season of calving, milk components and conformation class). Results from analyses of relative operating characteristics indicated that artificial neural networks could discriminate between mastitic states with an overall accuracy of 86%. This discriminatory ability was subject to patterns that existed in the training data files but was not affected by differing proportions of mastitic records. However, differing proportions of mastitic records had some effect on the particular purpose of the artificial neural network being developed: training with a higher proportion of mastitic cases increased the ability of an artificial neural network in discriminating positive from negative cases. Similar effects were obtained by modifying the threshold value used to categorize the ANN output, which constitutes a much simpler approach than modifying the proportion of cases in training data sets. Additional variables had little effect on the prediction accuracy, but this lack of effect needs to be verified for optimal artificial neural network configuration, data preprocessing, and new sources of information.

Computerized information systems can potentially help the dairy producer to deal with the increased complexity of decision making and availability of information in dairy farming. These systems, however, should be fully integrated to ensure a coordinated execution of dairy farming activities. A framework was developed, therefore, to support the creation of computerized management and control systems in dairy farming. Within this framework, a management and control system was defined as a network consisting of the management and control activities and the information flows that are involved in dairy farming. The management and control activities consist of a cycle of decision making, implementation, and assessment. These activities were classified according to level (strategic, tactical, operational, or regulatory) and sphere (breeding, health, nutrition, environment, milk production, fixed assets, labor, or finance). These activities can be performed by human beings or automated systems that are located on or off the farm. A large amount of information exchange exists among these management and control activities, between the overall management and control system, and between the physical farm environment and external agents. The interdependence among decisions at the various levels and spheres necessitates computerized management and control systems that are integrated and that support the exchange of information. The developed framework should facilitate the creation of such systems and could also act as a reference base for the analysis and improvement of existing dairy farm information systems.

A multi-network system, comprising a neural classifier and two specialized neural predictors, was developed for prediction of milk yield from monthly records of Holstein dairy cattle. The classifier categorized records based on high- (≥9000 kg) and low-yields (

Dairy producers receive large amounts of data from Dairy Herd Improvement agencies on a monthly basis, not only for the purpose of supplying summary statistics, but also as an aid in the detection of management weaknesses. The latter task can be carried out more efficiently by producers and their advisors if aided by decision-support systems, the role of which is to do preliminary data processing. Two software programs were developed to carry out such pre-processing for individual cows in terms of milk yield and persistency. Specifically, these programs analyze both traits as a function of parity, stage of lactation, and herd average production level, and are based on fuzzy sets, which furnish qualitative assessments of the deviations from standard values. This approach has proved to work well, and may be extended to other factors such as peak yield, body condition score and somatic cell count.

The effect of data preprocessing on the learning ability of artificial neural networks was investigated with regard to the impact of distributing the input vectors uniformly with respect to the output categories in the training data set. The analyses were performed for neural networks dedicated to (1) dairy cow culling classification and (2) milk yield prediction. The two types of neural network used for culling classification were backpropagation and learning vector quantization. For yield prediction, backpropagation was used. The study was repeated with several architectures for both types of network. Preprocessing of data did not have a large impact on the general performance of the networks, but did affect the results for each output category. The effects were more pronounced in the categories containing less frequent events, for which the results always improved. For the categories with larger number of records, balancing the data degraded the results. The respective improvements and degradation of the results occurred for both prediction and classification, with the two types of neural networks, and with all architectures tested. However, the magnitude of the effects varied with the type of neural network and with the architecture. The results of this study indicate that, in general, the distribution of outputs influences the learning process of neural networks for both types of application. The results also suggest that the types of output distribution required for the training of neural nets may depend on the specifics of each problem.

Since the main reason for disposal of dairy cows is low milk yield, implementation of an optimum selection program requires the prediction of cow performance with regard to production. The prediction of fat and protein content in milk are also rapidly becoming important factors for decisions related to breeding and herd policy. While, on average, traditional lactation models furnish good results, some improvement is possible when predicting the yield for an individual cow early in lactation. Artificial neural networks (ANNs), known to perform well in pattern recognition, may constitute an effective alternative to the traditional models. The objective of this research was to investigate how ANNs might be used to predict total milk, fat, and protein production for individual cows. Results indicated that ANNs generally performed at least as well overall as the model currently used by Canadian milk recording agencies, especially in the first third of lactation. This has important implications for early identification of superior animals. Predictions from both methods were relatively similar for the later stages of lactation. The addition of nontraditional data inputs such as average milk herd production and weight of cow improved the quality of prediction. Three different techniques were used to analyze the sensitivity of the ANN to different inputs, and their relative abilities are discussed. Results illustrate the potential effectiveness of ANNs in predicting milk yield and its composition and appear to justify further pursuit of this research.

A methodology was developed for estimating the parameters involved in a first-order autoregressive process; these parameters comprise a variance component associated with the random effect, a correlation coefficient, ρ, and a residual variance. These parameters were estimated using REML with an expectation-maximization algorithm. For two single-trait analyses (milk and fat production being the dependent variable), the example chosen for the analyses was year-month—treated as random and following a first-order autoregressive process—within fixed herd. Initially, estimates failed to converge, possibly because of a time trend in the data, which was not accounted for by the model. After the random effect that follows the first-order autoregressive process was redefined as month within fixed herd-year, the parameters converged, and ρ was estimated as .8 for milk and fat yield. Results suggest that the estimation procedures may be useful for situations when a first-order autoregressive process seems appropriate.

A first-order autoregressive process is proposed for modeling certain random effects in instances for which a more general model would involve too many (co)variance components. This process requires only two parameters, and its inverse—needed for its inclusion in mixed model methodology—is tridiagonal and easy to obtain. The (co)variance matrix for effects that follow a first-order autoregressive structure is introduced, and the rules for obtaining its inverse are derived and outlined in an algorithm (an example is also given). Incorporation of such a structure into the mixed model equations is also discussed, and an iterative procedure for obtaining solutions to this potentially large system of equations is outlined.

(Co)variance components for regional production of milk, fat, and fat percentage were estimated via multiple-trait REML using DHIA data from California, New York, and Wisconsin. Data consisted of registered and grade first lactation yields of Holstein cattle from 1970 through 1984. Records were limited to daughters from sires common to the pairs of states analyzed. Averaged numbers of records from California, New York, and Wisconsin were 419,185, 460,296, and 449,866. Genetic correlations between New York and Wisconsin for milk, fat, and fat percentage (.99, .98, and .99) were larger than those between California and New York (.95, .95, and .98) and those between California and Wisconsin (.94, .93, and .98). Heritabilities in New York for milk, fat, and fat percentage (approximately .27, .26, and .48) were larger than those of Wisconsin (approximately .23, .22, and .43) and California (approximately .22, .21, and .34). Grade records were also investigated; numbers of records in California were reduced by about 25%, whereas records in both New York and Wisconsin were reduced by about 66%. Genetic correlations were virtually unchanged, but heritabilities for milk and fat in the three states were similar (approximately .17) and significantly smaller than those from combined registered and grade daughters. In conclusion, there appeared to be no sire by region interaction.

Replacement heifers tend to calve substantially later than recommended in the province of Quebec in Canada. To support heifer management, large amounts of heifer growth data were collected over the past decade. The goal of this study was to explore the application of knowledge discovery techniques to these data to determine factors of interest associated with delayed age at first calving. Data covering heifer growth measurements, first lactation performance, type classification, genetics, and herd average milk recording information were collected and combined into a single animal level data set with 49,000 records. Twenty attributes of interest were derived from the available raw data and used for further analysis in addition to 22 existing attributes. Attribute values were cleaned by setting limits to the range of possible values. Statistical analyses and information visualization revealed that, on average, animals achieved recommended growth rates. Many animals were bred at a relatively heavy weight even if this resulted in delayed age at first calving. Most of the research effort was required for data preparation tasks. Information visualization and decision tree induction helped to identify interesting patterns associated with the relative weight at breeding.

Machine learning was employed to develop knowledge–based modules to automate the filtering of test–day data for group–average lactation curve analysis. The classification task consisted of identifying an individual cow’s milk–yield test as being either outlier or non–outlier. Experiments were carried out to explore the importance of deriving predictive attributes and choice of machine–learning technique. Data consisted of 1080 milk yield tests, of which 108 had been classified as outliers by a dairy nutrition specialist and 972 cases had been classified as non–outliers. Two different approaches to machine learning were applied to these data: decision–tree induction and naive–Bayes classification. Performance of the classifiers was estimated through ten–fold cross–validation, while relative operating characteristic curves were used to visualize the achieved trade–off between sensitivity and specificity. Use of an initial set of derived attributes significantly improved the performance compared to limiting attributes to those available to the domain specialist. However, adding even more complex attributes did not necessarily improve the classification performance. Overall, the naive–Bayes approach showed significantly better performance than decision–tree induction. For each machine–learning approach, three final classifiers, associated with a low, medium, and high filtering intensity, were generated from the 1080 cases. The expected true positive rate varied from 41% to 78% for false positive rates between 1.1% and 4.6%. However, due to the low prevalence of outlier tests, this performance was associated with a large number of false positives. The domain specialist considered the final classifiers of both approaches as plausible, but found the decision trees easier to understand than the naive–Bayes classifiers. Machine learning was considered to be a promising approach to assist knowledge acquisition.

Type classification records of Canadian Holsteins were investigated for evidence of heterogeneous variance across herds. Data consisted of records for 1,139,104 cows from 20,226 herds with classifications on 26 conformation traits collected from 1982 through 1992 and 338,046 cows from 9600 herds with classifications on 2 additional traits from 1990 through 1992. Phenotypic standard deviations of herd-round-classifier were fitted to a mixed model that included round, classifier, and region as fixed effects, herd size as a covariant, and herd as a random effect. Estimates of the variance components, solutions of fixed effects, and BLUP estimates for herd were obtained by maximum likelihood procedures. Repeatability of within-herd standard deviation across rounds ranged from 1.4 to 10.3% for the 28 traits. Type classification data were subsequently standardized for phenotypic standard deviations of herd-round-classifier that were derived from estimates of the fixed effects and the BLUP estimate of the herd effect. Genetic evaluations for cows and bulls were produced from adjusted and unadjusted data. Correlations between 3754 sire and 1,142,782 cow estimated transmitting abilities obtained from unadjusted and adjusted data were essentially unity. Although some evidence of heterogeneous variance existed across herds for 28 conformation traits, standardization of the classification records had only a minor effect on genetic evaluations.

1) In a beechwood, male shoots of Mercurialis perennis were more frequent than female shoots in canopy openings, while females were more frequent at treebases than in canopy openings. 2) Flowering shoots were more frequent than non-flowering in openings. 3) The pH, moisture and organic content of the soil were similar at treebases and openings. The light climate was similar in both habitats in April but greater in the openings in August. 4) Removal of canopies of Pteridium aquilinum and Rubus fruticosus, above populations of M. perennis, resulted in a greater increase in the numbers of male shoots than of female shoots. 5) These results suggest that male and female plants differ in numbers, or growth, in different woodland microhabitats and that these differences are correlated with the above-ground environment (e.g. light).

Heritabilities and genetic correlations were estimated for milk and fat yields and fat percentage from 305-d, mature-equivalent records of registered and grade Holsteins in California, New York, and Wisconsin. Parameter estimates were obtained with daughter on dam regression within herd-year-season and sire of daughter. Data were pairs of first lactation records (38,115 in California, 171,555 in New York, and 136,031 in Wisconsin) from 1975 through 1984. California had higher means for milk (9046kg) and fat (323kg) than the other two states (approximately 7840kg and 289kg), but heritabilities of .36 and .34 were not greatly different from those for New York (.34 and .33) or Wisconsin (.38 and .35). Heritability of fat percentage, however, was smaller for California (.55 vs .66 for New York and .68 for Wisconsin). Logarithmic and square root transformations of the data resulted in little change in heritability estimates. For California, genetic correlation between milk and fat yields (.70) was larger than for New York (.55) or Wisconsin (.56), and correlation between fat yield and fat percentage was smaller (.31 vs .45 and .41). For each state, estimates of heritabilities were larger for registered than grade pairs.