Churn Prediction in Internet Service Provider Companies

As a basic standard of life, internet connects millions of computers in a global network. People use, participate, or access the internet with the help of internet service providers (ISPs). To have better quality of connection, customers are prone to change their ISPs. In the competitive environment, ISPs endeavor to prevent losing their customers which are referred as churn. Thus, churn management takes an important place for ISPs. To investigate customer loyalty status, behavior, and information of the churn possibility in Turkey, a questionnaire is implemented. By using a real data obtained from a survey, promising and applicable results are obtained to predict the churn behavior of ISP customers in Turkey. As an extension of the study, the questionnaire will be applied for a larger population to find accurate results about churn situations. This study will help ISP companies to determine the required advertising campaigns for the customers.

Review of Business Intelligence and Intelligent Systems in Healthcare Domain

Business intelligence and corresponding intelligent components and tools have been one of those instruments that receive significant attention from health community. In order to raise more awareness on the potentials of business intelligence and intelligent systems, this paper aims to provide an overview of business intelligence in healthcare context by specifically focusing on the applications of intelligent systems. This study reviewed the current applications into three main categories and presented some important findings of that research in a systematic manner. The literature is wide with respect to the applications of business intelligence covering the issues from health management and policy related topics to more operational and tactical ones such as disease treatment, diagnostics, and hospital management. The discussions made in this article can also facilitate the researchers in that area to generate a research agenda for future work in applied health science, particularly within the context of health management and policy and health analytics.

A Soft Computing Approach to Customer Segmentation

There are a number of traditional models designed to segment customers, however none of them have the ability to establish non-strict customer segments. One crucial area that can meet this requirement is known as soft computing. Although there have been studies related to the usage of soft computing techniques for segmentation, they are not based on the effective two-stage methodology. The aim of this study is to propose a two-stage segmentation model based on soft computing using the purchasing behaviours of customers in a data mining framework and to make a comparison of the proposed model with a traditional two-stage segmentation model. Segmentation was performed via neuro-fuzzy two stage-clustering approach for a secondary data set, which included more than 300,000 unique customer records, from a UK retail company. The findings indicated that the model provided stronger insights and has greater managerial implications in comparison with the traditional two-stage method with respect to six segmentation effectiveness indicators.

On the Comparison of Quantitative Predictabilities of Different Financial Instruments

Huge amount of liquidity flows into a number of financial instruments such as stocks, commodities, currencies, futures, and so on every day. Investment decisions are mainly based on predicting the future movements of the instrument(s) in question. However, high frequency financial data are somewhat hard to model or predict. It would be valuable information for the investor if he or she knew which financial instruments were quantitatively more predictable. The data used in the model consisted of intraday frequencies covering the period between 1993 and 2013. An Artificial Neural Network model using Radial Basis Functions containing only past data of three different types of instruments (stocks, currencies, and commodities) to predict future high values on six different frequencies was applied. A total of 72 different artificial neural networks representing 12 different instruments were trained five times each, and their prediction performances were recorded on average. Considerably clear distinctions were observed on prediction performances of different financial instruments.

A Brief Review of Metaheuristics for Document or Text Clustering

Document clustering, which involves concepts from the fields of information retrieval, automatic topic extraction, natural language processing, and machine learning, is one of the most popular research areas in data mining. Due to the large amount of information in electronic form, fast and high-quality cluster analysis plays an important role in helping users to effectively navigate, summarize and organise this information for useful data. There are a number of techniques in the literature, which efficiently provide solutions for document clustering. However, during the last decade, researchers started to use metaheuristic algorithms for the document clustering problem because of the limitations of the existing traditional clustering algorithms. In this chapter, the authors will give a brief review of various research papers that present the area of document or text clustering approaches with different metaheuristic algorithms.

Image Mining

In today's World, huge multi-media databases have become evident due to the fact that Internet usage has reached at a very-high level via various types of smart devices. Both willingness to come into prominence commercially and to increase the quality of services in leading areas such as education, health, security and transportation imply querying on those huge multi-media databases. It is clear that description-based querying is almost impossible on such a big unstructured data. Image mining has emerged to that end as a multi-disciplinary field of research which provides example-based querying on image databases. Image mining allows a wide variety of image retrieval and image matching applications intensely required for certain sectors including production, marketing, medicine and web publishing by combining the classical data mining techniques with the implementations of underlying fields such as computer vision, image processing, pattern recognition, machine learning and artificial intelligence.

Swarm Intelligence Approaches to Shelf Space Allocation Problem with Linear Profit Function

The efficient management of shelf space carries critical importance on both the reduction of operational costs and improvement of financial performance. In this context, which products to display among the available products (assortment decision), how much shelf space to allocate the displayed products (allocation decision) and which shelves to display of each product (location decision) can be defined as main problems of shelf space management. In this paper, allocation problem of shelf space management is examined. To this end, a model which includes linear profit function is used for the shelf space allocation decision. Then, heuristic approaches are developed based on particle swarm optimization and artificial bee colony for this model. Finally, the performance analysis of these approaches is realized with problem instances including different number of products and shelves. Experimental results show that the proposed swarm intelligence approaches are superior to Yang's heuristics for the shelf space allocation model.

An Integrated Grey Relations Analysis and VIKOR Method for Multi Criteria Decision Making under Fuzzy Environment

Today, suppliers of companies are no longer local. Companies have to offer their products to the market just in time and as fast as possible in order to compete. This situation is possible by establishing an effective supply chain for the goods and services they need in the manufacturing system. Finding the right suppliers who are able to provide the companies with the high quality products and services at the reasonable price, at just on time and in the right quantities is an important issue concerned in the process of supply chains concept. There are certain techniques developed in this respect. Some of such methods are approaches developed for situations unmindful of fuzziness and vagueness. Nonetheless, the process of supplier selection contains both vagueness and fuzziness. This study improves the Grey Relational Analysis and VIKOR methods, to fuzzy and ambiguous environments. Then, these approaches are applied to a supplier selection problem, which is previously solved through fuzzy logic and AHP method in literature, and the comparative results of both techniques are given.

Use of Chaotic Randomness Numbers

The timeless search for optimizing the demand and supply of any resource is one of the main issues for humanity nearly from the beginning of time. The relevant cost of adding an extra resource reacts by means of more energy requirement, more emissions, interaction with policies and market status makes is even more complicated. Optimization of demand and supply is the key to successfully solve the problem. There are various optimization algorithms in the literature and most of them uses various algorithms of iteration and some degree of randomness to find the optimum solution. Most of the metaheuristic and artificial intelligence algorithms require the randomness where to make a new decision to go forward. So this chapter is about the possible use of chaotic random numbers in the metaheuristic and artificial intelligence algorithms that requires random numbers. The authors only provide the necessary information about the algorithms instead of providing full detailed explanation of the subjects assuming the readers already have theoretical basic information.

Data Mining for Multicriteria Single Facility Location Problems

This chapter focuses on available data analysis and data mining techniques to find the optimal location of the Multicriteria Single Facility Location Problem (MSFLP) at diverse business settings. Solving for the optimal of an MSFLP, there exists numerous multicriteria decision analysis techniques. Mainstream models are mentioned in this chapter, while presenting a general classification of the MSFLP and its framework. Besides, topics from machine learning with respect to decision analysis are covered: Unsupervised Principal Components Analysis ranking (PCA-rank) and supervised Support Vector Machines ranking (SVM-rank). This chapter proposes a data mining perspective for the multicriteria single facility location problem and proposes a new approach to the facility location problem with the combination of the PCA-rank and ranking SVMs.