Using POS Data for Price Promotions Evaluation

Price promotions have been largely dealt with in the literature. Yet there are just a few generalizations made so far about this powerful marketing communication tool. The obvious effect, that all authors who have studied price promotions emphasize, is quantity increase during price promotions. Inference studies about the decomposition of the sales promotion bump do not converge to a generalization or a law, but end in radically different results. Most of these studies use consumer panel data, rich of demographical characteristics and consumers’ purchasing history. Companies that use such data, available from marketing research industry, usually complain that data is old and expensive. The authors start with literature review on price promotions in which they present existing models based on consumer panel data (Bell, et al., 1999; Mela, et al., 1998; Moriarty, 1985; Walters, 1991; Yeshin, 2006). Next they present existing POS analysis models and compare their findings to show the high level of heterogeneity among results. All existing models are based on powerful databases provided by professional research institutions (i.e. Nielsen or IRI) that usually cover the whole market for the analysed brand category geographically. The authors next apply existing models to find which best suits data available for Slovenian FMCG market. They show two models analysis – quantity (SCAN*PRO) and market share (MCI) and their power for explanatory and forecasting research using POS data. Having dealt with more than 30 brand categories within a wider research, they conclude that the models developed are usable for a fast decision making process within a company, but their exploratory power is still poor compared to panel data.

Social Aid Fraud Detection System and Poverty Map Model Suggestion Based on Data Mining for Social Risk Mitigation

After last global financial crisis, one of the most important concerns of the governments became unemployment. Higher unemployment rates haves been forcing governments to develop some policies. Some of these policies has been included financial policies while some of them included social policies. One of the most important concerns of social policies is social risk mitigation and fight against poverty and social aids as its extensions. In general, measurement of social events have been mostly based on subjective statements. More specifically, targeting mechanisms have been using for determination of potential social aid owners. Most popular targeting mechanisms are subjective ones as well. In this chapter, an objective targeting mechanism model and a fraud detection system model have been developed via data mining for social aids as an identifier of poverty levels which includes early warning signals for inappropriate applications. Then, these models have been used for development of a poverty map. Developed new targeting mechanism which has been based on rating approach will be an alternative to Means Test and Proxy Means Test. In addition, social aid fraud detection system will be updated automatic with Intelligent System property and the poverty map computation approach can be used for absence of detailed data. Furthermore, Millenium Development Goals, Targeting Mechanisms, Poverty and Poverty Maps concepts have been reviewed from an analytical and objective point of view.

Data Mining Used for Analyzing the Bankruptcy Risk of the Romanian SMEs

The small and medium enterprises (SMEs) represent the backbone of the economy, playing a major economic and social role in the process of developing a dynamic economy. But the recent evolutions in the financial markets, the international financial crisis, the increased competition on markets, the lack of financial resources and the insufficient adaptation of many firms to the requests of the European market are new threats which can determine the bankruptcies of the Romanian SMEs. In this context, starting from the necessity to design an early warning system, we will elaborate a new model for analysis of bankruptcy risk for the Romanian SMEs that combine two main categories of indicators: financial ratios and non-financial indicators. The authors‘ analysis is based on data mining techniques (CHAID) in order to identify the firms’ categories accordingly to the bankruptcy risk levels. Through the proposed analysis model they try to offer a real surveillance system for the Romanian SMEs which can allow an early signal regarding the bankruptcy risk.

Designing an Early Warning System for Stock Market Crashes by Using ANFIS

This chapter focuses on building a financial early warning system (EWS) to predict stock market crashes by using stock market volatility and rising stock prices. The relation of stock market volatility with stock market crashes is analyzed empirically. Also, Istanbul Stock Exchange (ISE) national 100 index data used to achieve better results from the view point of modeling purpose. A risk indicator of stock market crash is computed to predict crashes and to give an early warning signal. Various data mining classifiers are compared to obtain the best practical solution for the financial early warning system. Adaptive neuro fuzzy inference system (ANFIS) model was proposed to forecast stock market crashes efficiently. Also, ANFIS was explained in detail as a training tool for the EWS. The empirical results show that the fuzzy inference system has advantages to gain successful results for financial crashes.

Black-Necked Swans and Active Risk Management

This chapter gives a review of recent developments in sequential surveillance and modeling of default probabilities of corporate and retail loans, and relates them to the development of early warning or quick detection systems for managing the risk associated with the so-called “black swans” or their close relatives, the black-necked swans.

On the Nature and Scales of Statistical Estimations Divergence and its Linkage with Statistical Learning

Besides the well-known commonplace, and sometimes also simply fantastic reasons for the existing breaks in the estimations of one and the same phenomena, substitution of concepts, manipulations, intentional distortions, all possible manipulations and frank lie there are their own technological reasons in the statistics for the similar breaks, which are being generated by some sort of circumstances of insurmountable force, which one should differ from well-known posy reasons, and therefore to consider in a special order. Predetermined objectively by conditioned divergence of the theoretical and empirical distributions, gaps between a nature and phenomenon, shape and its content, word and deed, these reasons (different from subjective reasons), limited by the extreme possibilities of human existence, can be overcome through the expansion of humans knowledge’s, which assumes reconsideration of the very basis of the modern science. Below we present some of the approaches towards such a reconsideration, which opens possibilities for the reduction of the huge gaps in modern statistical estimations of the same phenomena and its linkage with statistical learning.

Data Mining and Explorative Multivariate Data Analysis for Customer Satisfaction Study

By the early 1990s, the term “data mining” had come to mean the process of finding information in large data sets. In the framework of the Total Quality Management, earlier studies have suggested that enterprises could harness the predictive power of Learning Management System (LMS) data to develop reporting tools that identify at-risk customers/consumers and allow for more timely interventions (Macfadyen & Dawson, 2009). The Learning Management System data and the subsequent Customer Interaction System data can help to provide “early warning system data” for risk detection in enterprises. This chapter confirms and extends this proposition by providing data from an international research project investigating on customer satisfaction in services to persons of public utility, like education, training services and health care services, by means of explorative multivariate data analysis tools as Ordered Multiple Correspondence Analysis, Boosting regression, Partial Least Squares regression and its generalizations.

Bankruptcy Prediction by Supervised Machine Learning Techniques

It is very important for financial institutions which are capable of accurately predicting business failure. In literature, numbers of bankruptcy prediction models have been developed based on statistical and machine learning techniques. In particular, many machine learning techniques, such as neural networks, decision trees, etc. have shown better prediction performances than statistical ones. However, advanced machine learning techniques, such as classifier ensembles and stacked generalization have not been fully examined and compared in terms of their bankruptcy prediction performances. The aim of this chapter is to compare two different machine learning techniques, one statistical approach, two types of classifier ensembles, and three stacked generalization classifiers over three related datasets. The experimental results show that classifier ensembles by weighted voting perform the best in term of predication accuracy. On the other hand, for Type II errors on average stacked generalization and single classifiers perform better than classifier ensembles.

Data Mining and Economic Crime Risk Management

Economic crime is a billion dollar business and is substantially present in our current society. Both researchers and practitioners have gone into this problem by looking for ways of fraud mitigation. Data mining is often called in this context. In this chapter, the application of data mining in the field of economic crime, or corporate fraud, is discussed. The classification external versus internal fraud is explained and the major types of fraud within these classifications will be given. Aside from explaining these classifications, some numbers and statistics are provided. After this thorough introduction into fraud, an academic literature review concerning data mining in combination with fraud is given, along with the current solutions for corporate fraud in business practice. At the end, a current state of data mining applications within the field of economic crime, both in the academic world and in business practice, is given.

Collaborative Video Surveillance for Distributed Visual Data Mining of Potential Risk and Crime Detection

Thanks to fast technology advancement of micro-electronics, wired/wireless networks and computer computations in past few years, the development of intelligent, versatile and complicated video-based surveillance systems has been very active in both research and industry to effectively enhance safety and security. In this chapter, the authors first introduce the generations of video surveillance systems and their applications in potential risk and crime detection. For effectively supporting early warning system of potential risk and crime (which is load-heavy and time-critical), both collaborative video surveillance and distributed visual data mining are necessary. Moreover, as the surveillance video and data for safety and security are very important for all kinds of risk and crime detection, the system is required not only to data protection of the message transmission over Internet, but also to further provide reliable transmission to preserve the visual quality-of-service (QoS). As cloud computing, users do not need to own the physical infrastructure, platform, or software. They consume resources as a service, where Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS), Software-as-a-Service (SaaS), and pay only for resources that they use. Therefore, the design and implementation of an effective communication model is very important to this application system.