Search

• Supervised, univariate and entropy-based discretization strategies are presented. • The proposed discretizer outperforms other discretizers for non-monotonic attributes. • It is also able to capture monotonic relationships, making it flexible. • Discretization improves classifier accuracy with respect to the original data. Discretization is one of the important pre-processing steps for supervised learning. Discretizing attributes helps to simplify the data and make it easier to understand and analyze by reducing the number of values. It can provide a better representation of knowledge and thus help improve the accuracy of a classifier. However, to minimize the information loss, it is important to consider the characteristics of the data. Most approaches assume that the values of a continuous attribute are monotone with respect to the probability of belonging to a particular class. In other words, it is assumed that increasing or decreasing the value of the attribute leads to a proportional increase or decrease in the classification score. This assumption may not always be valid for all attributes of data. In this study, we present entropy-based, flexible discretization strategies capable of capturing the non-monotonicity of the attribute values. The algorithm can adjust the number of cut points and values depending on the characteristics of the data. It does not require setting of any hyper-parameter or threshold. Extensive experiments on different datasets have shown that the proposed discretizers significantly improve the performance of classifiers, especially on complex and high-dimensional data sets. [Display omitted] [ABSTRACT FROM AUTHOR]

This study presents a new linear multivariate decision tree (MDT) algorithm that incorporates linear programming and components of the branch-and-bound methodology, such as bound-based pruning and node selection strategies. MDTs are known to better separate classes than univariate counterparts. However, the methodology used to construct splitting hyperplanes significantly affects the tree's performance. To determine the splitting hyperplane at each node, we efficiently solve a deviation model, a linear program (LP) that minimizes the external deviation of misclassified alternatives from the splitting hyperplane. The proposed MDT, called the MDT-DevM, uses upper and lower bounds for pruning. The deviation at a node provides an upper bound for its child. The lower bound is derived from the leaf nodes. The proposed pruning rule much contributes to the induction of a shallower and more accurate decision tree. Several node selection strategies (depth_first , breadth_first and best_first) are tried in the proposed MDT. We have also presented a variant of MDT, called the MDT-SVM, which utilizes linear SVM to find the splitting hyperplane. We performed comprehensive experiments on artificial and real datasets with different levels of complexity. We compared the proposed methods with benchmark classifiers. The proposed MDTs outperform the linear classifiers. The MDT-DevM shows better performance as the data becomes more dispersed and imbalanced. • A new linear MDT with bounding and search strategies is presented. • The proposed bound-based pruning rule is more effective than traditional pruning rules. • The proposed MDT produces a shallower and more accurate decision tree. • It outperforms or is comparable to its univariate and multivariate counterparts. [ABSTRACT FROM AUTHOR]

A decision table is a practical tool that helps systems planners to make operational decisions, especially when they are under stress. With the effect of recent trends, such as the use of machine learning, data mining, and reinforcement learning methods, the maintenance decision has been a dynamic issue depending on system conditions. An expert may execute the maintenance or wait for the next periodic maintenance due to lack of maintenance workers, tools or budget, resources, etc., although the intelligent method predicts a failure approaching. Even sometimes, he/she may ignore the current periodic maintenance. Our method allows making some changes in the maintenance plan systematically. It integrates the results of preventive and predictive maintenance policies, and as different from the literature, it allows ignoring some maintenance actions depending on the maintenance resource levels in a decision table. Such a strategy helps to allocate limited resources to maintenance actions reasonably. We conducted an extensive simulation study on a real-life dataset. The preventive maintenance period is determined using classical approaches such as Weibull analysis. A machine learning algorithm is utilized to predict the type of failure. We have analyzed the performance of the proposed decision table approach under a variety of scenarios and with different parameter settings. We also showed the effect of parameter settings and the marginal utility of each maintenance policy. In addition, the approach provides several choices for planners. As a result, the proposed approach improves the system's sustainability compared to traditional policies. [ABSTRACT FROM AUTHOR]

One of the alternative investment strategies of large-sized enterprises is to improve the supplier's processes. Large-sized enterprises can increase the efficiency of own processes by investing on the manufacturing processes of suppliers, working with them continually. In this study, the selection problem of suppliers having the priority for investment is discussed with respect to different evaluation criteria. Analytical Network Process (ANP) method is used to find the weights of the criteria. The score of each supplier is determined by using the criteria weights from ANP in the Prometheell and Prometheell+Tch methods. However, considering the investing company, the total amount of turnover of suppliers from the company is also important. Since we considered that the turnover criterion is very dominant, we didn't include it in our analysis initially. The best supplier(s) is(are) determined by using the turnover criterion together with other criteria in the data envelopment analysis. The best supplier to be chosen is the one, which can be adapted to the process improvement program fastest. By investing on supplier's process, it would be taken under control better, and the improvement would be possible. In the application stage, the suppliers of an enterprise, which operates in the furniture industry, are evaluated according to nine criteria with the suggested method, and results are analyzed. [ABSTRACT FROM AUTHOR]

The multiobjective set covering problem (MOSCP), a challenging combinatorial optimization problem, has received limited attention in the literature. This paper presents a heuristic algorithm to approximate the Pareto set of the MOSCP. The proposed algorithm applies a local branching approach on a tree structure and is enhanced with a node exploration strategy specially developed for the MOSCP. The main idea is to partition the search region into smaller subregions based on the neighbors of a reference solution and then to explore each subregion for the Pareto points of the MOSCP. Numerical experiments for instances with two, three and four objectives set covering problems are reported. Results on a performance comparison with benchmark algorithms from the literature are also included and show that the new algorithm is competitive and performs best on some instances. [ABSTRACT FROM AUTHOR]