Search

Detecting changes of behaviors or events is crucial when updating existing knowledge in a dynamic business environment. Currently, data analysts can immediately collect data and easily access existing knowledge. However, that knowledge can also rapidly become outdated. This study discusses a form of knowledge, classifiable sequential patterns ( CSP s), defined as s → c , where s is a temporal sequence; c is a class label; and “→” is a sign which implies the sequential relationships between s (cause) and c (effect). If the CSP evolves into another, and the new knowledge is not updated, decision-makers would continue to work with the obsolete CSP . To the authors' knowledge, no study has addressed the topic of change mining in CSPs. To address this research gap, this study proposes a novel change-mining model, SeqClassChange , to identify changes in CSPs. Experiments were conducted with a real-world dataset to evaluate the proposed model.

Purpose Customer lifetime value (CLV) scoring is highly effective when applied to marketing databases. Some researchers have extended the traditional association rule problem by associating a weight with each item in a transaction. However, studies of association rule mining have considered the relative benefits or significance of “items” rather than “transactions” belonging to different customers. Because not all customers are financially attractive to firms, it is crucial that their profitability be determined and that transactions be weighted according to CLV. This study aims to discover association rules from the CLV perspective. Design/methodology/approach This study extended the traditional association rule problem by allowing the association of CLV weight with a transaction to reflect the interest and intensity of customer values. Furthermore, the authors proposed a new algorithm, frequent itemsets of CLV weight (FICLV), to discover frequent itemsets from CLV-weighted transactions. Findings Experimental results from the survey data indicate that the proposed FICLV algorithm can discover valuable frequent itemsets. Moreover, the frequent itemsets identified using the FICLV algorithm outperform those discovered through conventional approaches for predicting customer purchasing itemsets in the coming period. Originality/value This study is the first to introduce the optimum approach for discovering frequent itemsets from transactions through considering CLV.

This study proposes a model to detect the change in students' learning performance.This model helps educators to understand learning behaviors in dynamic environments.Experiments are conducted with real-life datasets to evaluate the effectiveness of the model. In recent years, as information technology has become more prevalent, learning management systems have arisen around e-learning and web-based platforms. As a result, huge quantities of data about students' learning process have been recorded and stored. Teachers can apply data-mining techniques to mine students' learning performance. One such technique is association rule mining, which can find correlations between student characteristics and performance. For instance, a rule (Attendance=Middle)?(Gender=Male)?(Semester=Low) indicates that the semester grade of students is at the Low level if their gender is Male and attendance rate is Middle, where Low and Middle are predetermined linguistic terms given by teachers. Teachers can rely on such rules to formulate their teaching strategies. However, these rules may be varied in different semesters because of the change of student characteristics or teaching method of teachers. The above rule is used to describe student behavior during the last semester, yet, within this semester, the rule changes to (Attendance=Low)?(Gender=Female)?(Semester=Low). Without updating this knowledge, teachers might adopt inappropriate teaching strategies for students who are learning in different ways across different semesters. In this study, we propose a new change mining model to discover the change in student learning performance and characteristics on the basis of association rules. We conducted experiments with real-life datasets to evaluate the effectiveness of the proposed model.

In real-life circumstances, managers often have to face various decision-making problems. Among them, the topic of multi-criteria decision making (MCDM) is one of the most important and complicated problems in the decision-making field. In general, the element of MCDM can consist of three parts: input, output, and the solution approach. The input is expressed as m alternatives with n criteria. It usually assumes that the underlying input data can be represented in a decision table. Correspondingly, the output is an optimal outcome, or a set of outcomes, resolved by different kinds of solution approaches. In this study, we propose a summarization technique to display the decision table as a new summarization table; allowing managers to make decisions more quickly. The proposed technique is to optimize a summarization result so that the degree of information lost is minimum. Since seeking for a minimum result is an NP-hard problem, we applied a genetic algorithm to improve the summarization result. In real-life circumstances, managers have to face various decision-making problems.MCDM is one of the most important problems in the decision-making field.The input data of MCDM can be represented in a decision table.We propose a summarization technique to display it as a new summarization table.

The needs of long-term care have increased substantially worldwide.We evaluate the system usage performance and continued use intentions of LTCIS.The result demonstrates that our proposed model explains 62.7% of the variance. The needs of long-term care (LTC) have increased substantially worldwide. In particular, Taiwanese society has experienced an increasing aging population because of the sharp decline in the birth rate and advanced medical technology. Some LTC facilities have introduced information technologies to satisfy the substantial demand. Compared with the high level of informatization observed at Taiwanese hospitals, LTC facilities still exhibit low informatization levels despite belonging to the same industry. Therefore, this study evaluated the system use performance and continued use intentions of the long-term care information system (LTCIS) by applying the well-known theories of the task-technology fit, system satisfaction, and postacceptance continuance models used in information systems area. The results showed that users' assessment of whether the LTCIS fulfilled their work and task needs primarily depended on the system quality, locatability of data, timeliness, ease of use, and system-user relationship. These factors also exerted a crucial influence on system use performance and user satisfaction. Moreover, the system use performance and user satisfaction further affected intentions of continued use.

While improving decisional quality is important to businesses, continued use of DMTs is a critical issue for managerial personnel. This problem mainly concerns the willingness of an individual to participate in the behavior. It can be further defined in a routine-based working behavior. This problem essentially involves three key issues, task fit, technology use, and habit. This study therefore integrates task-technology fit (TTF) model, expectation–confirmation mode (ECM), and habit, to examine the determinants of continued use of DMTs. Prior studies have focused on intention to use DMTs in the first time and only considered part of the three issues for identifying the determinants. 285 respondents from managerial personnel were collected to empirically evaluate this research model. The three issues are all important in influencing continuance use intention of DMTs. In particular, the task-technology fit indicates a direct effect on two factors of the technology use issue, user satisfaction and perceived usefulness, and an indirect effect on continuance use intention. User satisfaction and perceived usefulness, and habit are the key predictors of continuance use intention.

Traditionally, clustering is the task of dividing objects into homogeneous clusters based on their degrees of similarity. As objects are assigned to clusters, users need to manually give descriptions for all clusters. Characterizing clusters by hand can consume a great deal of time of users. In addition, users sometimes have no specific idea as to how to explain the clustering results; thus, they might give inappropriate descriptions. A clustering technique is proposed to discover conjecturable rules, providing descriptions of clusters with a decision tree classification technique. Every cluster in a conjecturable tree is depicted by only one conjecturable rule. However, less-utilized rules are not necessarily trivial. In some real-life circumstances, there might be some clusters which can be depicted by two or more rules, namely, recessive conjecturable rules. For example, customers usually prefer to buy inexpensive red wines; however, on certain occasions, such for a birthday celebration, they will buy expensive wine. Therefore, we know that there are some people who generally belong to a low-value cluster but may simultaneously be assigned to a high-value one. In this study, we propose a new discovery model for mining conjecturable rules to reveal this type of knowledge. The experimental results show that our proposed model is able to discover conjecturable rules as well as recessive rules. The results of sensitivity analysis are also given for practitioners' reference.

In today's business environment, there is tremendous interest in the mining of interesting patterns for superior decision making. Although many successful customer relationship management (CRM) applications have been developed based on sequential pattern mining techniques, they basically assume that the importance of each customer is the same. Previous studies in CRM show that not all customers make the same contribution to a business, and it is indispensible to evaluate customer value before developing effective marketing strategies. Therefore, this study includes the concepts of recency, frequency, and monetary (RFM) analysis in the sequential pattern mining process. For a given subsequence, each customer sequence contributes its own recency, frequency, and monetary scores to represent customer importance. An efficient algorithm is developed to discover sequential patterns with high recency, frequency, and monetary scores. Empirical results show that the proposed method is efficient and can effectively discover more valuable patterns than conventional frequent pattern mining.

The problem of mining partial periodic patterns is an important issue with many applications. Previous studies to find these patterns encounter efficiency and effectiveness problem. The efficiency problem is that most previous methods were proposed to find frequent partial periodic patterns by extending the well-known Apriori-like algorithm. However, these methods generate many candidate partial periodic patterns to calculate the patterns' supports, spending much time for discovering patterns. The effective problem is that only one minimum support threshold is set to find frequent partial periodic patterns but the results is not practical for real-world. In real-life circumstances, some rare or specific events may occur with lower frequencies but their occurrences may offer some vital information to be referred in decision making. If the minimum support is set too high, the associations between events along with higher and lower frequencies cannot be evaluated so that significant knowledge will be ignored. In this study, an algorithm to overcome these two problems has been proposed to generating redundant candidate patterns and setting only one minimum support threshold. The algorithm greatly improves the efficiency and effectiveness. First, it eliminates the need to generate numerous candidate partial periodic patterns thus reducing database scanning. Second, the minimum support threshold of each event can be specified based in its real-life occurring frequency.

With advances in information and network technologies, lots of data have been digitized to reveal information for users by the construction of Web sites. Unfortunately, they are both overloading and overlapping in Internet so that users cannot distinguish their quality. To address this issue in education, Hwang, Huang, and Tseng proposed a group decision system to evaluate the quality of educational Web sites by users' and experts' opinions. Their investigative source is solely stemmed from human intention, called the subjective perspective, to make judgments on the quality of Web sites. However, the nature of human beings in making decisions has a gap between intention and behavior. Asking people for eliciting thought is arduous to cause this gap. Human behavior, namely the objective perspective, is the other essential source to obtain human thinking and real doings. For this reason, we can use data mining approaches to acquire the objective source. In this research, we propose an integrated decision model applied in evaluating educational Web sites from the fuzzy subjective and objective perspectives. The former source is extracted by inquiring human opinion using a questionnaire, while the latter is gained automatically by a data mining technique, fuzzy clustering. An empirical study is carried out to validate the model capability.

Sequential pattern mining is essential in many applications, including computational biology, consumer behavior analysis, web log analysis, etc. Although sequential patterns can tell us what items are frequently to be purchased together and in what order, they cannot provide information about the time span between items for decision support. Previous studies dealing with this problem either set time constraints to restrict the patterns discovered or define time-intervals between two successive items to provide time information. Accordingly, the first approach falls short in providing clear time-interval information while the second cannot discover time-interval information between two non-successive items in a sequential pattern. To provide more time-related knowledge, we define a new variant of time-interval sequential patterns, called multi-time-interval sequential patterns, which can reveal the time-intervals between all pairs of items in a pattern. Accordingly, we develop two efficient algorithms, called the MI-Apriori and MI-PrefixSpan algorithms, to solve this problem. The experimental results show that the MI-PrefixSpan algorithm is faster than the MI-Apriori algorithm, but the MI-Apriori algorithm has better scalability in long sequence data.

Items sold in a store can usually be organized into a concept hierarchy according to a taxonomy. Based on the hierarchy, sequential patterns can be found not only at the leaf nodes (individual items) of the hierarchy, but also at higher levels of the hierarchy; this is called multiple-level sequential pattern mining. In previous research, taxonomies had crisp relationships between the categories in one level and the categories in another level. In real life, however, crisp taxonomies cannot handle the uncertainties and fuzziness inherent in the relationships among items and categories. For example, the book Alice's Adventures in Wonderland can be classified into the Children's Literature category, but can also be related to the Action &Adventure category. To deal with the fuzzy nature of taxonomy, we apply fuzzy set techniques to concept taxonomies so that the relationships from one level to another can be represented by a value between 0 and 1. Accordingly, a fuzzy multiple-level mining algorithm, the fuzzy multi-level sequential mining algorithm (FMSM), is proposed to extract fuzzy multiple-level sequential patterns from databases. In addition, another algorithm, named the CROSS-FMSM algorithm, is developed to discover fuzzy cross-level sequential patterns. Experiments using synthetic datasets show the algorithms' computational efficiency and scalability, and a real dataset is used to prove the patterns' effectiveness.

Given a sequence database and minimum support threshold, the goal of mining quantitative sequential patterns is to discover the complete set of sequential patterns with purchased quantities in databases. Although this type of pattern can provide more information than the traditional sequential pattern, it also causes a sharp boundary problem. This means that when an item's quantity is close to the boundary of two adjacent quantity intervals, it is either ignored or overemphasized. In view of this weakness, a recent paper from Hong, Kuo, and Chi proposed a new kind of extended patterns, called fuzzy quantitative sequential patterns (FQSP), where an item's quantity in the pattern is represented by a fuzzy term rather than a quantity interval. In their work an Apriori-like algorithm was developed to mine all FQSP. In this paper, we propose a new and novel algorithm to mine FQSP based on the divide-and-conquer strategy. Since the proposed algorithm greatly reduces the candidate subsequence generation efforts, the performance is improved significantly. Experiments show that the proposed algorithm is much more efficient and scalable than the previous algorithm.

Given a sequence database and minimum support threshold, the task of sequential pattern mining is to discover the complete set of sequential patterns in databases. From the discovered sequential patterns, we can know what items are frequently brought together and in what order they appear. However, they cannot tell us the time gaps between successive items in patterns. Accordingly, Chen et al. have proposed a generalization of sequential patterns, called time-interval sequential patterns, which reveals not only the order of items, but also the time intervals between successive items . An example of time-interval sequential pattern has a form like (A, I/sub 2/, B, I/sub 1/, C), meaning that we buy A first, then after an interval of I/sub 2/ we buy B, and finally after an interval of I/sub 1/ we buy C, where I/sub 2/ and I/sub 1/ are predetermined time ranges. Although this new type of pattern can alleviate the above concern, it causes the sharp boundary problem. That is, when a time interval is near the boundary of two predetermined time ranges, we either ignore or overemphasize it. Therefore, this paper uses the concept of fuzzy sets to extend the original research so that fuzzy time-interval sequential patterns are discovered from databases. Two efficient algorithms, the fuzzy time interval (FTI)-Apriori algorithm and the FTI-PrefixSpan algorithm, are developed for mining fuzzy time-interval sequential patterns. In our simulation results, we find that the second algorithm outperforms the first one, not only in computing time but also in scalability with respect to various parameters.

The aggregation of individuals’ preferences into a consensus ranking is a decision support problem which has been widely used in various applications, such as decision support systems, voting systems, and recommendation systems. Especially when applying recommendation systems in business, customers ask for more suggestions about purchasing products or services because the tremendous amount of information available can be overwhelming. Therefore, we have to gather more preferences from recommenders and aggregate them to gain consensuses. For an example of the preference ranking, C>A≥D≥B indicates that C is favorable to A, and A (D) is somewhat favorable but not fully favorable to D (B), where > and ≥ are comparators, and A, B, C, and D are items. This shows the ranking relationship between items. However, no studies, to the best of our knowledge, have ever developed a recommendation system to suggest a temporal relationship between items. That is, “item A could occur during the duration of item B” or “item C could occur before item D”. This type of recommendation can be applied to the reading order of books, course plans in colleges, or the order of taking medicine for patients. In this study, we propose a novel recommendation model to discover closed consensus temporal patterns, where closed means that the patterns are only the maximum consensus sequences.

A learning management system (LSM) prevails and it accumulates an amount of data about the progress of student learning, demographic and students’ background over different sessions. Educators are very concerned about the shifts of unknown relationships among a thousand variables about students in a LMS for adjusting their teaching strategies and pedagogies. However, educators are not satisfied with the traditional reports, which are explored with limited relative variables on the issue of summative assessment and learning achievement in static sessions. The information about the changes of unknown relationships among many variables cannot be produced with statistical methods in traditional reports. Our study proposes a mining change of fuzzy quantitative association rules model to reveal the information. This model can discover the six types of changes rules from unknown relationships among many variables with nominal or numerical attributes. Experiments are carried out to evaluate the proposed model. We empirical demonstrate how the model helps educators understand the changing characteristics of students and to modify their teaching practices.