In this paper, we focused on developing efficient mining algorithm for multiclass classification from large of collection imbalanced data. And gives well sorted data. In the field of data mining, classification techniques can be used to find various selective features. This paper presents an innovative and efficient classification technique which includes the processes of feature selection and map reduction, to improve the effectiveness of using data for finding relevant and interesting information. In proposed system we can take sufficient .txt file as inputs & we apply variance algorithm & generate expected results. Classification is method to perform on poorly & minority class examples when the dataset is extremely imbalanced.