Paper Title
Effective Classification Using A Dataset Based On Recursive Logical And Analytical Data
Abstract
Data classification is defined as the classification of data using related attributes and the algorithms such as
decision tree, K-nn Algorithm etc.[1][3][5][7] The system is based on producing a fast and optimal method to classify a
Stock Market dataset and help identify the most acceptable record. It is based on the Logical analysis of data where a dataset
is classified based on various constraints to find a suitable record. The database is first cleansed to remove the redundant data
which provides the dataset. The dataset consisting of company’s maximum investment ratio and quarter wise profit and loss
percentage is then put through a process of discretization, where the data is converted into a binary form (also called as
binarization) for an efficient process. The discretized data is then provided as a dataset which is used for further
classification. The decision tree algorithm is applied where the number of investment ratios is taken as input. The weight and
coverage problem is used to split the dataset into patterns by cut points. The given constraints act as the cut points to split
the dataset. The K-nN algorithm groups the nearest data, in this case, the dataset with nearest possible profit percentage and
investment ratio. From the developed pattern, the classified data is provided to the client. The system increases the efficiency
by reducing the number of choices to select from.
Keywords— Decision tree , K-nn, LAD and Recursive LAD.