In the information age, enterprises’ activity of any organizational-legal form is accompanied by recording and registration of all data of their activities. In connection with improving technologies for recording and storage of data, huge amounts of information have afflicted people in various areas. This remarkable flow of data is continuously gathered in the process of commercial, industrial, medical, or scientific activities. Powerful computer systems storing and managing huge databases became an integral part of the daily life of organizations and their personnel. Data mining and knowledge discovery in database (KDD) are the methods of detection of necessary information for decision-making and actions towards sustainable development of any enterprise. This report aims at investigation of data mining and KDD concepts as well as the necessity of their application in the retail industry.
Most organizations in the course of their activities gain huge amounts of data, but the only thing they want to receive from them is necessary information. KDD is a process of detecting useful information from large amount of data. This knowledge may be presented as patterns, rules, forecasts, as well as linkages between the elements of data. The key instrument of KDD is analytical technologies of data mining implementing such tasks as classification, clustering, regression, forecasting, prediction and others (Sumathi & Sivanandam, 2006).
However, in accordance with the concept of KDD, the effective process of knowledge discovery is not limited to its analyzing. KDD consists of the sequence of operations needed for support of the analytical process. These include consolidation of data, preparation for data analysis, clearance of the factors inhibited correct analysis, data optimization, analysis using the methods of data mining, and interpretation of results and their implementation in business applications (Triantaphyllou & Felici, 2006).
A preparatory stage of original data set aims at creating data selection including various sources. For that purpose, there must be developed tools for access to various data sources. The support of data warehouses and the presence of the semantic layer allow using not technical terms but business concepts in preparing of data source. In order to effectively apply the methods of data mining, it is necessary to pay serious attention to the issues of data pre-processing. Data can contain gaps, noises, abnormal values, as well as be excessive or insufficient. Data should be of high quality and correct, so the first important step of KDD is data pre-processing. The step of transformation and normalization of data is necessary to put information for further analysis. Besides, some methods of analysis require the source data to be in a particular form only. For example, neural networks work only with normalized numeric data. At the stage of data mining, there are different algorithms that are used for knowledge discovery such as neural networks, decision trees, algorithms of clustering, and the establishment of associations. The final step is post-processing, which includes interpretation of results and implementation of knowledge in business applications.
Data mining is the main step of KDD process. It is a set of approaches and methods for extracting data that includes search for patterns and dependence between the data. The positive feature of data mining is a singularity of wanted patterns and hidden knowledge. According to one of the founders of this direction, Gregory Piatetsky-Shapiro, data mining is the process of search of unknown, nontrivial, and operationally useful information in the raw data for its further use in various spheres of human activity (Frawley, Piatetsky-Shapiro, & Matheus, 1991).
Original writing according to your instructions
Deadlines from 3 hours to 60 days
All disciplines covered
Skilled writers with Master’s/PhD degrees
Personal data security
Instant replies to all your questions
Data mining is a complex of numerous different methods of knowledge discovery. The basis of data mining is mathematical tools arising and developing on the basis of the achievements of applied statistics, pattern recognition, methods of artificial intelligence, and database theory. The choice of method often depends on the type of available data and what kind of information is needed to get.
From the marketing point of view, it is important to identify the pattern of associative type, based on the research in the supermarket, it allows to find that, for example, 65% of customers who buy chips also purchase drinks. If there is a chain of time related events, the methods of data mining reveal the pattern of sequence type, for instance, the purchase of a new computer in 75% of cases leads to buying a new software. The detection of patterns of clustering type allows defining the features that characterize the group and its homogeneity, which is important in the selection and evaluation of the target audience.
Data mining is of great value for managers and analysts in their daily activities, because it allows obtaining tangible benefits in the competition. In some areas of the business, large companies cannot compete with small firms owing to their individual approach to clients based on thoroughly researched preferences.
Book The Best Top Expert at Top-papers.com
Your order will be assigned to the most experienced writer in the relevant discipline. The highly demanded expert, one of our top-10 writers with the highest rate among the customersHire a TOP Writer for 10.95 USD
Therefore, organizations record all information about regular clients and from clients with the OLTP system, then the data from various systems are selected for storage and analyzing, and on the basis of this analysis, a company makes decisions and actions that are useful for business.
Presently, retail companies collect detailed information about each individual purchase using the credit card and computerized control systems. There are three common tasks that can be solved using data mining in the retail industry. First of all, the analysis of shopping cart (analysis of similarities) is designed to identify items that customers tend to purchase together. Knowledge of shopping cart is needed to improve advertising and a strategy for the creation of reserves of goods and their shelf space in the trading floors (Pal & Jain, 2007). The second task is the investigation of temporal patterns, since it helps retailers to make decisions about creating inventory. It gives the answers to the question: if a customer has bought a basic good today, how long after that he would buy other related products? Finally, the last challenge is a creation of predictive models that allow trading enterprises to understand the needs of different categories of customers with a certain behavior, for example, customers who buy products from well-known designers or attending the sales. This knowledge is needed to develop precisely targeted, cost-effective interventions for the promotion of products (Alkhalifa, 2010).
KDD helps to interpret the results and to apply this knowledge in business applications. For example, a trading company has a network of shops and needs to get sales forecasts for the next month. The first step is to collect the history of sales in each store and combine it in the total data sample. The next step is preprocessing of the collected data such as grouping the goods by months or development of sale curves and factors influencing the amount of sales. Next, one constructs a model of correlation between sales volumes and various factors. Defining this predictive value, it can be used, for example, in applications to optimize better placement of goods in warehouses.
Data mining methods and KDD are widely used in business applications by analysts and senior executives. These methods allow solving difficult practical issues without special economic and mathematical training. The relevance of using data mining and KDD in business relates to fierce competition (Maimon & Rokach, 2006).
Forecasting is one of the most common tasks of data mining and KDD. In particular, planning and budgeting require forecasting the sales volumes and other parameters taking into account numerous interrelated factors, whether seasonal, regional, economic, or others.
Data mining and KDD methods can also provide a marketing analysis that is one of the main elements in the success. To develop an effective marketing plan, one must, therefore, know how such factors as the cost of goods, cost of product promotion and advertising affect the sales level. Neural network models allow managers and analysts to predict such influence.
The third task is the analysis of staff work. The productivity of employees depends on their level of training, work experience, relationship with management, and many other factors. In analyzing the influence of these factors, it is possible to develop a methodology to increase productivity and to propose the optimal strategy of recruitment in the future.
Bibliography / Reference Page
Get all these features for $80 FREE
If a company distributes advertising and product samples by mail, it makes sense to analyze the effectiveness of such activity. It is possible to identify the range of potential buyers and to evaluate the likelihood of their purchase. Besides, such trading company can also try various forms of correspondence and choose the most successful of them.
Data mining methods also allow profiling customers. Using neural network models among numerous clients of the firm, analysts can choose those whose cooperation is most profitable, in other words, to get a portrait of a “typical customer.” Sales managers as well as banks often face the challenge of assessing the creditworthiness of customers.
The next task is evaluation of potential customers. Planning exploratory talks, one must determine how likely it will end with signing of a contract or products’ sales. The analysis of experience with customers allows identifying the characteristics of those applications that resulted in actual sales. Using the results of this analysis, managers can focus on more promising customer applications.
The last challenge of data mining and KDD is analysis of marketing research results. The interviewing of customers is needed to assess the reaction of customers on a company’s policy in the field of product distribution, pricing, and specifications of the products. This allows improving the decision-making process on prices and characteristics of products such as design, functionality, or packaging.
Exclusive savings! Save 25% on your ORDER
Get 15% OFF your FIRST ORDER + 10% OFF every order by receiving 300 words/page instead of 275 words/page
These types of tasks are relevant not only for retailers but for all business industries such as banking, insurance, financial markets, production and others. In the processes of working with large amounts of generated information, there is a need to monitor the development of the market over time, and this is virtually impossible without automation of the analytical activities (Maimon & Rokach, 2006).
To summarize, KDD as well as its most important stage data mining are the process of discovering useful information in large volumes of accumulated data. A scope of application of this method is not restricted; it is used in every area that is attached to the data. In particular, KDD is of great value for managers and analysts in their business activities. The use of data mining techniques can provide significant benefits in an increasingly competitive world.