K. Thippeswamy Book order

Retrieval and analysis of information from knowledge base is a basic and vital function that any system is expected to provide with the growing complexity of information. Various institutions/organizations generate valuable information in different domains which is queried and analyzed by users for their purposes. Transporting Reservation, Finance, E-Real estate, Life science, Pharmacy etc, are few of the knowledge driven sectors where data analysis is imperative. Most of the applications performing these tasks are predominantly database driven and tightly coupled with the database system. These limits are the possibilities of seamless database integration with other sources of knowledge and also their ability to create and make changes in the information structures.

According to certain criteria, the classes are identified by using classification techniques, which is considered as data mining tool. When compared with smaller class, the classification results (i.e., accuracy) for bigger class are deviating and the traditional classification procedures provides inaccurate results, which is known as Class Imbalance problem. A class is formed with unequal size, where this type of data is represented and combined as class imbalance data. There are two various categories are presents in class imbalance domain, namely minority (i.e., smaller) and majority (i.e., bigger) classes. The major aim of this research work is to identify the minority class accurately. In this research, two significant methodologies are proposed such as (i) Adaptive-Condensed Nearest Neighbor (ACNN)Algorithm, and (ii) Local Mahalanobis Distance Learning(LMDL) based ACNN algorithm. These methods are significantly improving the imbalanced data classification.