Search

In this paper we develop a new machine learning estimator for ordered choice models based on the Random Forest. The proposed Ordered Forest flexibly estimates the conditional choice probabilities while taking the ordering information explicitly into account. In addition to common machine learning estimators, it enables the estimation of marginal effects as well as conducting inference and thus provides the same output as classical econometric estimators. An extensive simulation study reveals a good predictive performance, particularly in settings with nonlinearities and high correlation among covariates. An empirical application contrasts the estimation of marginal effects and their standard errors with an Ordered Logit model. A software implementation of the Ordered Forest is provided both in R and Python in the package orf available on CRAN and PyPI , respectively. [ABSTRACT FROM AUTHOR]

There is a widely spread belief in the Bayesian network (BN) community that the overall accuracy of results of BN inference is not too sensitive to the precision of their parameters. We present the results of several experiments in which we put this belief to a test in the context of medical diagnostic models. We study the deterioration of accuracy under random symmetric noise but also biased noise that represents overconfidence and underconfidence of human experts.Our results demonstrate consistently, across all models studied, that while noise leads to deterioration of accuracy, small amounts of noise have minimal effect on the diagnostic accuracy of BN models. Overconfidence, common among human experts, appears to be safer than symmetric noise and much safer than underconfidence in terms of the resulting accuracy. Noise in medical laboratory results and disease nodes as well as in nodes forming the Markov blanket of the disease nodes has the largest effect on accuracy. In light of these results, knowledge engineers should moderately worry about the overall quality of the numerical parameters of BNs and direct their effort where it is most needed, as indicated by sensitivity analysis. [ABSTRACT FROM AUTHOR]

Currently, the tasks of intelligent data analysis in medicine are becoming increasingly common. Existing artificial intelligence tools provide high effectiveness in solving these tasks when analyzing sufficiently large datasets. However, when there is very little training data available, current machine learning methods do not ensure adequate classification accuracy or may even produce inadequate results. This paper presents an enhanced input-doubling method for classification tasks in the case of limited data analysis, achieved via expanding the number of independent attributes in the augmented dataset with probabilities of belonging to each class of the task. The authors have developed an algorithmic implementation of the improved method using two Naïve Bayes classifiers. The method was modeled on a small dataset for cardiovascular risk assessment. The authors explored two options for the combined use of Naïve Bayes classifiers at both stages of the method. It was found that using different methods at both stages potentially enhances the accuracy of the classification task. The results of the improved method were compared with a range of existing methods used for solving the task. It was demonstrated that the improved input-doubling method achieved the highest classification accuracy based on various performance indicators. [ABSTRACT FROM AUTHOR]

This article is a continuation of the work published in the previous issue of the journal.1 The main approaches to constructing scenarios for long-term socio-economic development are compared, both structural and morphological. The results of the scenario are verified using the traditional structural method based on the morphological approach, and a system of mutually agreed upon scenarios for long-term socio-economic and scientific and technological development is proposed (based on the structural approach). [ABSTRACT FROM AUTHOR]

North-Eastern (NE), India and its adjoining region is one of the sixth most seismically active regions of the world. In the present investigation, the return period of earthquake and probability of occurrence inferred from Gutenberg–Richter (GR) relation was estimated for NE, India region and its vicinity. When we consider the entire NE, India region and its vicinity, it evidently suggested that the return period of earthquakes of 7 ≤ Mw ≤ 8.6 is short, which ranges from 32.73 to 162.59 years. It was observed that the earthquake occurrence from infinitesimally short interval t~0 for Mw~3.6–4 is embedded with 100% probability. The earthquakes of Mw~4.1–5.3 reach 100% in 10 years. Similarly, Mw~5.4–5.7 reaches to 100% in 20 years. Likewise, Mw~5.8–5.9, 6.0–6.1 and 6.2 reach ~100% in 30, 40 and 50 years, respectively. For large earthquakes of Mw~7.0–8.0, the probability of occurrence reaches >80% in 100 years. This observation strongly indicates that the likelihood of earthquakes occurring in the north-eastern region of India and its surrounding areas tends to increase over time. Further, the region was divided into four zones, namely Block I (26.5–28.5ºN; 89–95ºE), Block II (26.5–28.5ºN; 95–97.5ºE), Block III (23–26.5ºN; 93–97.5ºE) and Block IV (23–26.5ºN; 89–93ºE) based on seismicity and the major tectonic domains of the region. In terms of return period based on GR-relation and stochastic observations, we may conclude that the risk associated with occurrence of earthquake is highest in Block IV, followed by Block III, Block I and Block II respectively. Further, a comparison of the probabilities of earthquake return period considering seismogenic depths along with hypocentral depth data for different blocks was investigated for a comprehensive understanding of seismic occurrences over time. However, overall, the patterns and trends observed remain consistent, emphasizing the seismic activity within each block and its associated return periods. The stochastic observations and findings are elaborately accentuated in the article. [ABSTRACT FROM AUTHOR]

Although familiar with the developments in probability theory of his time, Vilfredo Pareto made little use of this tool in his writings, preferring theoretical constructions based on experimentation and observation. This article attempts to reconstruct Pareto's overall approach to probability by examining his references to the distribution of income, an economic fact that lends itself to probabilistic investigation. The result of this research shows how Pareto alludes to the application of probability to income and social groups, but leaves the task to his followers. JEL classification B31, B4, C1. [ABSTRACT FROM AUTHOR]

The International Society for Animal Genetics (ISAG) currently advocates for a transition towards single nucleotide polymorphism (SNP) markers as a potential alternative for equine parentage verification. To ascertain the efficacy of this transition, it is imperative to evaluate the performance of parentage testing using SNPs in juxtaposition with short tandem repeats (STRs). As per ISAG's recommendation, we used an equine genotyping‐by‐sequencing panel with 144 SNPs for this purpose. Equine parentage is currently realized using 16 microsatellites (STRs) excluding the LEX3 marker. In this study, 1074 horses were genotyped using the 144 SNPs panel, including 432 foals, 414 mares, and 228 stallions, from five different breeds: 293 Arabians, 167 Barbs, 189 Thoroughbreds, 73 Anglo‐Arabians, and 352 Arabian‐Barbs. As a result, two SNPs markers were eliminated from the panel system due to inconsistent amplification across all examined individuals leaving 142 SNPs markers for analysis. A comparative analysis between SNPs and STRs markers revealed that the mean expected heterozygosity was 0.457 for SNPs and 0.76 for STRs, while the mean observed heterozygosity stood at 0.472 for SNPs and 0.72 for STRs. Furthermore, the probability of identity was calculated to be 5.722 × 10−57 for SNPs and 1.25 × 10−15 for STRs markers. In alignment with the Hardy–Weinberg equilibrium in polyploids test, 110 out of the total SNPs were consistent with the Hardy–Weinberg equilibrium in polyploids test (p > 0.05). Employing both SNPs and STRs markers, the mean polymorphic information content was discerned to be 0.351 for SNPs and 0.72 for STRs. The cumulative exclusion probabilities for SNP markers exceeded 99.99%, indicating that the 142 SNPs panel might be adequate for parentage testing. In contrast, when utilizing STRs markers, the combined average exclusion probabilities for one and both parents were determined to be 99.8% and 99.9%, respectively. Our comprehensive study underscores the potential of SNPs in equine parentage verification, especially when compared to STRs in terms of exclusion probabilities. As a corollary, the application of SNPs for parentage verification and identification can significantly contribute to the conservation initiative for the five Moroccan horse breeds. Nonetheless, further research is required to address and replace the deficient SNPs within the panel. [ABSTRACT FROM AUTHOR]

In the ever-evolving world of veterinary care, the occurrence of bone fractures in canines poses a common and complex problem, especially in extra-small breeds and dogs that are less than 1 year old. The objective of this research is to fill a gap in predicting the risk of canine bone fractures. A machine learning method using a random forest classifier was constructed. The algorithm was trained on a dataset consisting of 2261 cases that included several factors, such as canine age, gender, breed, and weight. The performance of the algorithm was assessed by examining its capacity to forecast the probability of fractures occurring. The findings of our study indicate that the tool has the capability to provide dependable predictions of fracture risk, consistent with our extensive dataset on fractures in canines. However, these results should be considered preliminary due to the limited sample size. This discovery is a crucial tool for veterinary practitioners, allowing them to take preventive measures to manage and prevent fractures. In conclusion, the implementation of this prediction tool has the potential to significantly transform the quality of care in the field of veterinary medicine by enabling the detection of patients at high risk, hence enabling the implementation of timely and customized preventive measures. [ABSTRACT FROM AUTHOR]

Gold prices have been of major interest for a lot of investors, analysts, and economists. Accordingly, a number of different modeling approaches have been used to forecast gold prices. In this manuscript, the geometric Brownian motion approach, used in the pricing of numerous types of assets, is used to forecast the prices of gold at yearly, monthly, and quarterly frequencies. This approach allows for simulating one-period-ahead prices and the associated probabilities. The expected prices obtained from the simulated prices and probabilities are found to provide reliable forecasts when compared with the observed yearly, monthly, and quarterly prices. [ABSTRACT FROM AUTHOR]

We identify and analyze three caveats that may arise when analyzing the linguistic abilities of Large Language Models. The problem of unlicensed generalizations refers to the danger of interpreting performance in one task as predictive of the models’ overall capabilities, based on the assumption that because a specific task performance is indicative of certain underlying capabilities in humans, the same association holds for models. The human-like paradox refers to the problem of lacking human comparisons, while at the same time attributing human-like abilities to the models. Last, the problem of double standards refers to the use of tasks and methodologies that either cannot be applied to humans or they are evaluated differently in models vs. humans. While we recognize the impressive linguistic abilities of LLMs, we conclude that specific claims about the models’ human-likeness in the grammatical domain are premature.

We identify and analyze three caveats that may arise when analyzing the linguistic abilities of Large Language Models. The problem of unlicensed generalizations refers to the danger of interpreting performance in one task as predictive of the models' overall capabilities, based on the assumption that because a specific task performance is indicative of certain underlying capabilities in humans, the same association holds for models. The human-like paradox refers to the problem of lacking human comparisons, while at the same time attributing human-like abilities to the models. Last, the problem of double standards refers to the use of tasks and methodologies that either cannot be applied to humans or they are evaluated differently in models vs. humans. While we recognize the impressive linguistic abilities of LLMs, we conclude that specific claims about the models' human-likeness in the grammatical domain are premature. [ABSTRACT FROM AUTHOR]

Denić (2018, 2019, To appear) observes that the availability of distributive inferences—for sentences with disjunction embedded in the scope of a universal quantifier—depends on the size of the domain quantified over as it relates to the number of disjuncts. Based on her observations, she argues that probabilistic considerations play a role in the computation of implicatures. In this paper we explore a different possibility. We argue for a modification of Denić's generalization, and provide an explanation that is based on intricate logical computations but is blind to probabilities. The explanation is based on the observation that when the domain size is no larger than the number of disjuncts, universal and existential alternatives are equivalent if distributive inferences are obtained. We argue that under such conditions a general ban on 'fatal competition' (Magri 2009a,b, Spector 2014) is activated, thereby predicting distributive inferences to be unavailable. [ABSTRACT FROM AUTHOR]

In the ever-evolving world of veterinary care, the occurrence of bone fractures in canines poses a common and complex problem, especially in extra-small breeds and dogs that are less than 1 year old. The objective of this research is to fill a gap in predicting the risk of canine bone fractures. A machine learning method using a random forest classifier was constructed. The algorithm was trained on a dataset consisting of 2261 cases that included several factors, such as canine age, gender, breed, and weight. The performance of the algorithm was assessed by examining its capacity to forecast the probability of fractures occurring. The findings of our study indicate that the tool has the capability to provide dependable predictions of fracture risk, consistent with our extensive dataset on fractures in canines. However, these results should be considered preliminary due to the limited sample size. This discovery is a crucial tool for veterinary practitioners, allowing them to take preventive measures to manage and prevent fractures. In conclusion, the implementation of this prediction tool has the potential to significantly transform the quality of care in the field of veterinary medicine by enabling the detection of patients at high risk, hence enabling the implementation of timely and customized preventive measures.

In this perspective review, we assess fundamental flaws in Darwinian evolution, including its modern versions. Fixed mutations 'explain' microevolution but not macroevolution including speciation events and the origination of all the major body plans of the Cambrian explosion. Complex, multifactorial change is required for speciation events and inevitably requires self-organization beyond what is accomplished by known mechanisms. The assembly of ribosomes and ATP synthase are specific examples. We propose their origin is a model for what is unexplained in biological evolution. Probability of evolution is modeled in Section 9 and values are absurdly improbable. Speciation and higher taxonomic changes become exponentially less probable as the number of required, genetically-based events increase. Also, the power required of the proposed selection mechanism (survival of the fittest) is nil for any biological advance requiring multiple changes, because they regularly occur in multiple generations (different genomes) and would not be selectively conserved by the concept survival of the fittest (a concept ultimately centered on the individual). Thus, survival of the fittest cannot 'explain' the origin of the millions of current and extinct species. We also focus on the inadequacies of laboratory chemistry to explain the complex, required biological self-organization seen in cells. We propose that a 'bioelectromagnetic' field/principle emerges in living cells. Synthesis by self-organization of massive molecular complexes involves biochemical responses to this emergent field/principle. There are ramifications for philosophy, science, and religion. Physics and mathematics must be more strongly integrated with biology and integration should receive dedicated funding with special emphasis for medical applications; treatment of cancer and genetic diseases are examples. [ABSTRACT FROM AUTHOR]

Mathematical education in Greece is constantly evolving in the pursuit of optimal learning outcomes for students despite their cognitive differences. This study seeks to gain insight into the use of new technologies in teaching probability theory and the gender differences in the comprehension of probability theory. To this end, a survey was conducted involving 500 students of the Department of Statistics and Insurance Science at the University of Piraeus. The respective questionnaire involves questions of self-reported results and employs the Likert scale to obtain the students' perceptions. Our data demonstrate no difference among the genders regarding the use of new technologies or their performance (i.e., the test scores) in the relevant courses. [ABSTRACT FROM AUTHOR]

The study uses numerical weather prediction models to predict the occurrence of heavy convective rainfall associated with the passage of the African Easterly Wave (AEW) during the period 23–26 August 2017 over Nigeria. Fraction skill score (FSS) and method for object‐based diagnostic evaluation (MODE) verification techniques were applied to verify how well the models predict the high‐impact event and to demonstrate how these tools can support operational forecasting. Ensemble model forecasts at a convective scale from UK Met Office Unified Model (MetUM) and a one‐way nested weather research and forecasting (WRF) model were compared with the integrated multisatellite retrievals for global precipitation measurement (IMERG GPM). The purpose is to examine skills of improved model resolution and ensemble in reproducing rainfall forecasts on useful scales and how the skill varies with spatial scale. WRF 2 and 6 km model forecasts show comparable skill at smaller grid scales. The skill of MetUM improves dramatically when the verification statistics are applied to the ensemble mean of the binary fields of the individual member forecast. The object‐based analysis reveals a similar structure as observed, although displaced eastwards. Most improvement occurred for heavier rainfall events associated with the passage of the AEW. WRF 6 km compares reasonably well with WRF 2 km in terms of shape and structure of rainfall underscoring the ability of the model to reasonably represent convection at 6 km horizontal resolution. The ensemble members in MetUM explicitly reproduce convection at 4 km resolution but are displaced at about 166 km behind observed rainfall. [ABSTRACT FROM AUTHOR]

This study seeks to identify the ontological and epistemological foundations of linguistic theory, in light of recent developments in scientific practice that takes language as its object. It requires a reconsideration of the existential concept of language and raises the issue of the existence of ontological and epistemological constants in linguistic theory. Taking into consideration the difference in the nature of linguistic sound and meaning. The approach adopted is historical and comparative, examining the major ontological and epistemological milestones. At the ontological level, the study notes that René Descartes linked the existence of language to mental existence, but the theory of relativity and statistical theory have transformed language into a virtual digital system, calling into question its ontological framework. Not to mention the fact that language is no longer merely a mental product, as evidenced by the existence of automatically generated linguistic texts. Epistemologically, the verification of linguistic knowledge has evolved, moving from logical proof to experimentation, and is now tied to utility and added value, with the virtual simulation of artificial intelligence as the environment for experimentation and proof. [ABSTRACT FROM AUTHOR]