Abstract
This study proposes two predictive models of classification that allow to identify, at the end of the 1st and 2nd semesters, the undergraduate students of a higher education institution more prone to academic dropout. The proposed methodology, which combines 3 popular data mining algorithms, such as random forest, support vector machines and artificial neural networks, in addition to contributing to predictive performance, allows to identify the main factors behind academic dropout. The empirical results show that it is possible to reduce to about 1/4 the 4 tens potential predictors of dropout, and show that there are essentially two predictors, concerning student’s curriculum context, that explain this propensity. This knowledge is useful for decision-makers to adopt the most appropriate strategic measures and decisions in order to reduce student dropout rates.

Abstract

Abstract
Build to Order or Make to Order is a common approach for highly configured products such as special vehicles (vehicles that are adapted and altered to suit a specific purpose). Examples of such vehicles are special ambulances as well as vehicles adapted for the support and transport of passengers with less mobility. In this type of business, operations are scheduled in response to a confirmed order received from a final customer. Thus, the variability and the uncertainty characterizing what is project based, generate a complexity that requires specifically tailored managerial approaches to handle all the involved processes - from design and engineering to production and delivery. Hence, in this accentuated complexity, it is extremely important to guarantee that both the material and information flows are efficient and effective. The present study, framed in a program of operational improvement in a manufacturer of special vehicles, aims to address some concrete improvement opportunities related to the significant number of raw materials stockouts and to the high number of changes made by the client after production has started. In fact, during the manufacturing and assembly process, there are constant changes that delay and difficult planning and consequently decreases the overall efficiency and effectiveness. Strategies to address all these matters are to be identified and applied. © IEOM Society International.

Abstract
It is imperative that online companies have a complete in-depth understanding of online behavior in order to provide a better service to their customers. This paper proposes a model for real-time basket addition in the e-grocery sector that includes predictors inferred from anonymous clickstream data, such as a Markov page view sequence discrimination value. This model aims at anticipating the addition and the non-addition of items to customers' market basket, in order to enable marketers to act conveniently, for example recommending more appropriate items. Two classification techniques are used in the empirical study: logistic regression and random forests. A real sample of anonymous clickstream data taken from the servers of a European e-retailing company is explored. The empirical results reveal the high predictive power of the model proposed, based on the explanatory variables introduced, as well as the supremacy of random forests over logistic regression.

Abstract
Overlap has been identified in previous works as a significant obstacle to automated diagnosis using data mining algorithms, since it makes it impossible to discern how each machine influences product quality. Several solutions that handle overlap have been proposed, but the final result is a list of potential overlapped root causes. The goal of this paper is to develop a solution resilient to overlap that can determine the true root cause from a list of possible root causes, when possible, and determine the conditions in which it is possible to identify the root causes. This allows for a better understanding of overlap, and enables the development of a fully automatic root cause analysis for manufacturing. To do so, we propose an automatic root cause analysis approach that uses causal inference and do calculus to determine the true root cause. The proposed approach was validated on simulated and real case-study data, and allowed for an estimation of the effect of a product passing through a certain machine while disregarding the effect of overlap, in certain conditions. The results were on par with the state-of-the-art solutions capable of handling overlap. The contributions of this paper are a graphical definition of overlap, the identification of the conditions in which is possible to overcome the effect of overlap, and a solution that can present a single true root cause when such conditions are met.

Abstract
To improve manufacturing processes, it is essential to find the root causes of occurring problems, in order to solve them permanently. Automatic Root Cause Analysis (ARCA) solutions aid analysts in finding such root causes, by using automatic data analysis to improve the digital decision. When trying to locate the root cause of a problem in a manufacturing process, a phenomenon can occur that disrupts the application of ARCA solutions. Overlap, as we denominated, is a phenomenon where local synchronicities in the manufacturing process lead to data where it is impossible to discern the influence of each location in the quality of products, which impedes automated diagnosis, especially when using classifiers. This paper identifies and defines overlap, and proposes a two-phase ARCA solution that uses factor-ranking algorithms, instead of classifiers. The proposed solution is evaluated in simulated and real case-study data. Results proved the presence of overlap in the datasets, and its negative impact on classifiers. The proposed solution has a positive performance detecting root causes even in the presence of overlap.