First an application for twitter is necessary to get user details as key consumers, consumer secret, access token and key token access .Proceeding further, we have to write a code for retrieve the JSON object of Twitter tweets using the keyword that we seek. Here in our project, the keyword will be the name of parties contesting the elections as BJP, AAP, Congress, etc. Collected data will be stored in the temporary table using JSON .The collected data is stored in MS Excel. 2. The data collected has then been preprocessed to remove unnecessary data. Various preprocessing steps such as stop words, usertag, hashtag removal, casefolding, etc have been applied for data cleaning. 3. The words and emoticons expressing any kind of sentiment have been replaced with keywords. These keywords are used as features. 4. The next step is Feature Extraction. A list of ten features has been extracted for each tweet. 5. The extracted features are then normalised in the range of 0 to 1. 6. Cross correlation has been used as a method of Feature Reduction. One among the closely related features is removed. 7. Then, K-means, an unsupervised learning method has been applied to the remaining features to cluster the data into positive and negative classes. 8. Differential Evolution, an Optimization algorithm is then applied on the normalised feature set. 9. Accuracies of both algorithms is then calculated and compared to find out the more efficient algorithm.