K-Nearest Neighbor Query Processing using Utility Mining

Mining classification is the process of developing rules to group data tuples together based on certain features. Mining high utility itemsets from a transactional database refers to the process of discovery of items with high utility like profits. A large number of algorithms have been proposed for this problem, but it will generate a large number of itemsets. This will degrade the performance of the system. Inorder to overcome this problem here propose a method of utility mining. To perform the mining process, the system keeps a file called log file which contains the information’s regarding the previous searches of a particular user. This will help to identify the browsing behavior of the user. Here mining is done by classifying the items into promising and unpromising factors. A minimum utility value is assigned in order to classify the items. If the percentage count of a particular item chosen by the user is greater than minimum utility, then that item is considered as promising factor. The items with promising factor will be listed in the list. If the percentage count of a particular item chosen by a user is less than minimum utility, then hat item will be filtered out from the list. User will choose a particular item from the list. Selected item and the current location of the user will be transferred as a query. Both current location and selected item is processed inorder to find the nearest neighbor. Current location is found with the help of HTML5 geolocation technology. Nearest neighbors are found by comparing the longitude and latitude of the users current location with the set of items in the database. The nearest neighbors will be listed and user will chose a neighbor as per his requirement. The selected item will be located in the map. The system will also sow the shortest path, distance and time required to reach from source to destination. If the user selects a particular item then it will be considered as successful request and it will be stored in the log file. This will improve the performance of the system. By using his method more accurate results are obtained.