Foundation of Data Mining - Research Paper Example

? Data Mining Introduction Data mining, also known as knowledge discovery, is the process of extracting and analyzing data from different sources and summarizing it into helpful information. This information can be useful in increasing revenue, cutting cost of production, or both. Data mining software is a computer aided process of extracting and analyzing hidden predictive information from a large set of data (Hoptroff & Hoptroff, 2001). Data mining tools helps in predicting the behaviors and future trends of a business’ operations, thus allowing it make proactive and knowledge-based strategies. Data mining tools are vital because they have significantly reduced the time taken in answering business questions, which were traditionally too much time consuming to analyze. Currently, most organizations have adopted and implemented the existing data mining software and hardware platforms to improve the value of their stored data. These hardware platforms can be integrated with new products and system as technology advances. Integrating data mining hardware platforms with other parallel processing computers or high performance client/server improves the analysis of massive databases (Hoptroff & Hoptroff, 2001). Foundation of Data Mining Data mining techniques emerged as a result of product development and a long process of research. This idea was first developed when businesses began storing business information on computers. Significant improvements have been witnessed in data access and generated technologies, which allow users to search their data, in real time (Williams & Simoff, 2006). Data mining software is currently available for use, in the business world, because of the three technologies that support it, and they include data mining algorithms, massive data collection, and powerful multiprocessor computers (Williams & Simoff, 2006). The amount of raw data stored in business databases is currently exploding. A database is measured in gigabytes and terabytes. In the current, competitive business environment, raw data alone does not provide enough information for studying and predicting the market environment. This has called for the need to convert these terabytes of raw data into other significant insights that easily provide a guide for their investment, marketing and management strategies (Prabhu, 2004). Data Warehouses Significant improvements in data transmission, data capture, storage capabilities, and processing power are enabling companies to consolidate their various databases into data warehouse (Prabhu, 2004). Data warehousing is the process of centralizing data retrieval and data management. Data warehouses store large amounts of data based on certain categories that make data more easily to sort, retrieve, and interpret. They also enable managers and executives to manage a series of business transactions, and other information that help in making informed business decisions. Researchers have predicted that all companies shall have adopted and integrated data mining tools, in their business, by the year 2020 (Prabhu, 2004). Companies benefit from data mining when meaningful patterns and trends are extracted from the stored data. How Data Mining Works Data mining tools employ modeling as a technique for performing data analysis. Modeling involves the creation of a model in one situation that is known, and applying the results in another situation where the results are unknown (Kargupta, 2007). Computers are equipped with lots of information about a number of situations, whose answers are known. The data mining software, on the computer, runs through the data, and filters the aspects of data that match the designated model. Once the model is developed it can be applied in similar situations, whose answers are unknown. This technique has been in use over the past centuries, but it recently became applicable, in the business field, when communication and data storage capabilities required the collection and storage of huge amounts of data, and the ability to automate modeling techniques to compute data directly (Kargupta, 2007). Tasks of Data Mining A wide range of companies including health care, aerospace, manufacturing transportation, finance, and retail are already using data mining techniques and tools, although data mining technology is still being explored and improved. The use of mathematical and statistical techniques and pattern recognition technique to search through warehoused data helps data analysts recognize significant patterns, relationships, facts, trends, anomalies, and exceptions, which might otherwise go unnoticed (Williams & Simoff, 2006). In the business environment, data mining techniques and tools are used to discover the relationships and patterns, in data, to help in the formulation of better strategies and decisions. Therefore, data mining helps in the identification of sales trends, prediction of customer loyalty, and the development of better marketing strategies. Some of the other uses include market segmentation, customer churn, fraud detection, direct marketing, interacting marketing, marketing basket analysis, and trend analysis (Prabhu, 2004). Similarly, data mining tools can provide ideas of identifying new business opportunities through automated prediction of customers’ behaviors and trends, and automated identification of previously unknown patterns (Hoptroff & Hoptroff, 2001). Different organizations dig through large volumes of data, by using massively parallel computers, to establish a pattern about their products and customers. For instance, a grocery chains store that notices that women buying loaves of bread usually walk out with two packets of milk, can use this information to start up a store that brings these commodities under one roof. Data Mining Technologies Data mining employs mathematical algorithms and techniques as its analytical technique (Han et al. 2011). The use of graphical interfaces, which managers and executives use easily, has also led to the increased use of data mining tools. Some of the tools used include artificial neural networks and genetic algorithm. The former tool is a non-linear predictive model that resembles biological networks, in structure (Han et al. 2011). They are of two types: (1) decision tree, which generate rules for the classification of a data base, and rule induction, which enhances the extraction of useful information from a database. (2) Genetic algorithm is based on the concepts of mutation, natural selection, and genetic combination (Han et al. 2011). Real-World Examples An organization’s customer care department can dig into its customer-call database in order to manage its communications network effectively. It is possible to discover certain unmet customer needs through its data mining technology. The information gathered can be used to predict the additional services that should be incorporated in its communication network. The ability to discover changes, in customers’ needs and preferences, is a vital tool in giving a business a competitive advantage (Williams & Simoff, 2006). For instance, a financial institution looking for ways to increase its annual revenue from its credit card operations decided to test this non intuitive possibility. Will interest earned and usage of credit card increases if the institution lowers its minimum required payment? It then dug through terabyte of data representing average credit card balances, payment timelines, payment amounts, and credit limit usage among other key parameters over the past two years. It then developed a model that tests the effects of the proposed policy change on certain customer categories. This institution discovered that lowering minimum payment requirement, for a small customer category, significantly extended indebtedness and average balances periods (Williams & Simoff, 2006). The Future of Data mining In the short-term period, data mining techniques will help in raising profit margins, formulating new marketing techniques, and focusing advertisements on potential customers with new tastes and preferences. In the medium term, it can help in rooting out phone numbers of lost friends, finding best airfare, and finding the best prices on shares. In the long-term period, data mining may help medical researchers to develop new treatments and prevention for diseases, and determine the status of the natural environment (Kargupta, 2007). Privacy Concerns Most people and businesses are worried that data mining techniques reveal too much information about their personal details, thus interfering with their privacies. For instance, with data mining technology, it is possible to collect information about an individual’s telephone calls, employment application forms, flights, credit card records, warranty card send in, and every school record. If this information is collected and stored under one place, then it will take less than an hour in knowing a person’s personal details. Additionally, this information is available for everyone, in any part of the world, through the internet (Kargupta, 2007). References Han et al. (2011). Data Mining: Concepts and Techniques: Concepts and Techniques. New York: Elsevier. Hoptroff S. K, Hoptroff R. (2001). Data Mining and Business Intelligence: A Guide to Productivity. London: Idea Group Inc (IGI). Kargupta H. (2007). Data mining: next generation challenges and future directions. Michigan : The University of Michigan. Prabhu C. S. (2004). Data Warehousing: Concepts, Techniques, Products and Applications. New York: PHI Learning Pvt. Ltd. Williams G. J, Simoff S. J. (2006). Data Mining: Theory, Methodology, Techniques, and Applications. New York: Springer. Read More