Text data is a high-dimension and sparse data. The classification of text is the task of assigning a category or label to a given document. This is a supervised learning task, where a model is trained on a labeled dataset of text documents, and then used to predict the category or label of new, unseen documents. The common text classification tasks include sentiment analysis, spam detection, and news article categorization. The classification of unbalanced data remains a great challenge as the accuracy is not achieved to great extent. The different approach for handling the unbalanced nature of data is proposed which mainly focuses on the generation of synthetic data to balance the data artificially, its dimension reduction part, thus achieving good accuracy for classification of data using different machine learning techniques. With the help of SVM, authors are able to achieve 69.5% accuracy using Adasyn technique. The care of oversampling was done as it may lead to overfitting problems.