Analysis of internet access television user data by using data mining approach

The volume and variety of data increases day by day depending on technological developments and needs. It has become a necessity to collect and analyze the data accurately and discover the relationships between the data. Several data mining approaches such as classification, clustering and association rule are widely used to satisfy this needs. The aim of thesis study is to obtain information for creation of user-friendly, robust and segmented products as a result of the analysis of user data. By using clustering method and association rules, it is targetted to group users according to their usage habits. It is aimed to make customer segmentation through the clustering method. As a result of association rule analysis, it is aimed to determine which ports are used together for product design. It is also intended to evalute the reliability of examined product according to usage time. Within the scope of this thesis, the usage time and port usage data of internet-connected television users of a consumer electronic company were investigated. Analyzes were carried out by following the data mining process steps. The R language was used during the preprocessing, conversion, analysis steps of the data mining. 17 attributes related to usage habits were evaluated. K-means clustering method was used in the modeling phase for usage habits. Internal indexes were used to measure the clustering performance and determine the optimum number of clusters. As a result of the clustering study, it has been determined that users mostly prefer to use digital ports, especially external ports and satellite broadcasts. It has been determined which ports are frequently used in the relevant product segment. Association rule analysis was performed according to minimum support and confidence criteria determined. Association rules containing different numbers of groups were created. Utilization rates of ports that are identical to each other were evaluated. The reliability analysis technique was used for the life time evaluation of product. Changes in luminous flux over the years have been calculated. The luminous flux value at the 82,5 °C junction temperature which was calculated with the Arrhenius temperature life model parameters , was determined by means of the interpolation method. It has been predicted when LED flux will decrease to %70 level. By using the determined annual average watching time data, reliability and failure rate values were calculated. In this study, the Human Development Report data published by the United Nations Development Program was used. For the years between 1995 and 2021 , trends of human development level among 31 countries in the European continent were evaluated. The difference between the countries in the European continent in terms of the level of development was analyzed by the Anova method. Post-hoc analysis was conducted by using Games Howell test to determine which countries differ. Using the data of the human development report, the relationship between the television watching time and the level of human development was examined by correlation method. As a result of the correlation analysis, it was seen that there is a weak negative relationship between the level of human development and television viewing rates.