Data Mining and Behavior of Customers - Essay Example

? Data mining QUESTION Benefits of data mining process Predictive analytics to understand the behavior of customers The use of predictive analytics helps in the determination of a predictive score for the elements associated to the organization. The major organizational element, in this case, is the customers. The predictive scores inform the business about the most probable action by the customer. The production of predictive scores occurs when the subject organization design a predictive model. The predictive model work measures predictive scores based on the company’s data (Han et al, 2011). The predictive scores produced by the predictive analytics helps to increase the customer responses during the strategized marketing campaigns. The predictive score also helps in increasing the conversions and clicks, which in turn, help in decreasing the associated churns. Associations discovery in products sold to customers The dynamism of the market has led to the influence on the way the business interacts with their customers. The contemporary market bears no guarantee for the loyalty of a customer. This situation has led to the use of data mining in allowing for associations discovery on the goods sold to the customers. Association discovery consist of rules that use an antecedent (if) and a consequent (then) which represent items associated to the obtained customer’s data (Vaidya et al, 2006). The “if” is found in the data while “then” represent an item, which is in integration with the antecedent. The association rules works through analysing the if/then patterns yielded by the customers. The organization then records the patterns to help in identifying the most important relationships. The criteria used involve marking a either pattern as “support” or “confidence”. These criteria are important in studying the behavior of the consumer towards the products of any given organization. The association rules can be applicable in product clustering, catalog design and store layout (Han et al, 2011). Web mining to discover business intelligence from Web customers Web mining is an important application for data mining helping to study the web patterns. This application is important to organizations, which would like to discover then behavior of the web customers. Web mining works through gathering information from the websites using the traditional data mining whereby there is integration of data obtained from the interaction of the potential customers around the web. Web mining presents the business community with the ability to apply intelligent marketing strategies because of the knowledge about the customers. It helps the business to establish a close relationship with the website customers because of the satisfactory analysis on the market. Clustering to find related customer information Clustering in data mining is an important activity for businesses, which helps the latter to find various customers with similar taste to each other. It works through clustering data objects having same attributes concerning the market trend. Through clustering, the firms are able to select high-quality clusters that have low inter-cluster similarity and high intra-cluster similarity. This means that the more similarity between members of a cluster than to a member of another cluster. The process of selecting the clusters in the market is useful in situations where there are many cases with no grouping. The use of clustering algorithms helps in finding the natural groupings. QUESTION 2 Reliability of data mining algorithms The data mining algorithm reliability can be accessed through analysing how they create the data models from a given data. The algorithms are reliable because of the study of specific trends about a given data. The obtained result about the trend helps in identifying the optimal parameters required for the creation of data mining models. The presented optimal parameters helps in deducting appropriate patterns and statistics required for the given study. QUESTION 3 Privacy concerns Stereotypes Stereotypes can easily arise from the operation of KDDM when the criteria for studying the customers relied on race, gender or nationality. This concern is always very sensitive because the information obtained can lead to application of the intended commercial tools on the individuals based on their nationality or race. For instance, a bank may use data mining in studying pattern between two ethnic communities then apply their credit based on the attribute. The information obtained about a certain community can be used in discriminating them regardless of their contribution to the operation of the firm. Guarding personal data from KDDM researchers The need to find information about a given occurrence in the society requires the use of data mining in analysing the personal information related to situation. For instance, understanding an epidemic requires the use of data mining in analysing medical records of the individuals. This means that data collection by the KDDM researchers, despite being personal, helps in building knowledge in the society. However, the collection of personal data poses privacy threats because of the availability of the information in databases. The information can easily leak to the public databases. Combination of patterns Combination of patterns present privacy concerns in data mining process since it has both positive and negative implication. Combination of at least two patterns facilitates the disclosure of information about individuals. Analysing the presented data sets presents the personal information about a certain individual especially during the discovery about the data sets (Vaidya et al, 2006). For instance, when studying the spread of disease in a given community, it will be easy for the public to learn about a certain individual in the given data sets. The public can easily make conclusion about the data leading to the threat of privacy. Validity Stereotype concern is not valid for the data mining tools since it does not arise from the actual process. This is because Stereotypes in the data mining process arise from the perception possessed by the studied communities and not the technology used. Concern about combination of patterns is valid since the data obtained about the data sets can depict the information about a given subject. The public can easily make conclusion about the subject individual from the data set thereby leading to threat of privacy. Guarding data from the KDDM researchers is also valid since the information obtained help in studying a trend in the society, while also present privacy threats to the given subjects. Solution The major solution to the arising stereotypes during the data mining requires individual to understand that it arises from sociological domain. Stereotypes in the data mining process arise from the perception possessed by the studied communities and not the technology used. Consequently, the KDDM researchers try to alleviate this problem through creating awareness about the impact of stereotypes, on the studied community, on the data mining results. Guarding personal data from KDDM researchers is more of technological oriented because it relates to the ability of the KDDM researchers to obtain information about individuals. The possible way applied in allaying this issue is through removing identifiers such as names, addresses and telephone numbers (Nisbet et al, 2009). The removal of the identifiers helps in providing general information about a given community and not at personal levels. The major strategy put forward allaying combination patterns is the use of security control mechanism. The security control mechanism helps in providing quality statistics about a given set of data while preventing disclosure of confidential information. The security control mechanism has also helped in hiding individual values yielded in a given study hence preventing the concerns from combination of patterns. Reference Han, J., Kamber, M., & Pei, J. (2011). Data Mining: Concepts and Techniques. Burlington: Elsevier Science. Nisbet, R., Elder, J. F., & Miner, G. (2009). Handbook of statistical analysis and data mining applications. Burlington, MA: Academic Press/Elsevier. Vaidya, J., Zhu, M., & Clifton, C. W. (2006). Privacy preserving data mining. New York: Springer. Read More