Data Mining of data set from twitter ,

Mining Data set from Twitter for finding sentiments about different Telecommunication Companies , the task: . The tasks 1 and 2 have already been done 1. Downloading relevant tweets from Twitter using Twitter API. (Done) 2. Manually label the tweets as positive, negative and neutral or the students can use binary classification (positive or negative). (Done) 3. Using tools like NLTK (Natural Language Tool-Kit) or Rainbow [login to view URL]~mccallum/bow/rainbow/ to construct a document-word matrix while removing the stop words (e.g. a, an, the, is etc.) and performing stemming (converting run, runs, runner => run) at the same time. In our case, the documents would be tweets. 4. Apply different classification algorithms such as Decision Tree (ID3), Naïve Bayesian, Nearest Neighbor (kNN / IBK), Support Vector Machines (SVM), and Neural Networks (available sometimes as Multi-Layer Perceptron (MLP)). Try to use diverse algorithms. You could use WEKA / SPSS / R. 5. Analyze the results with respect to accuracy of different algorithms under different conditions. 6. Explain the results in simple English and with the help of graphs..