In this paper a hybrid algorithm for text data categorization has been proposed. There are four different classifiers of filtrations have been used. First classifier is numeric classifier (NC). NC is used for the numeric value classification and removal. Second classifier is separator classifier (SC). SC is used for the delimiter value classification and removal. Third classifier is all classifier (AC). AC is used for the grammar classification and removal. Last classifier is manual data classifier (MDC). MDC is used for the manual data value classification and removal. Then the associated cuckoo search optimization (CSO) based hierarchy optimization has been applied on the obtained data. Overall accuracy obtained by ourapproach is approximately 95%.

In this paper a survey and analysis based on topic based data classification has been presented. It includes the topic based data orientation, data categorization, document clustering, etc. This study provides the analytical way to analyze the methods previously published and provide explorative way of the approaches presented. It also provides the discussion based on the attributes and property used and explored. This study provides the discussion of different partitioning algorithm, different grouping methods and classification approaches.Based on the study, future enhancements have been suggested.