ranu010101 / twittersentimentanalysis

Sub Task B: Message Polarity Classification Given a message, classify whether the message is of positive, negative, or neutral sentiment. For messages conveying both a positive and negative sentiment, whichever is the stronger sentiment should be chosen.

Files Info:

1. ./dataset/<train/test>ingDatasetComplete.txt Contains complete <train/test>ing dataset in following format: "\t""\t""\t"

2. ./dataset/<train/test>ingDatasetProcessed.txt Contains processed <train/test>ing dataset without the entries for which tweet are not available in following format: "\t""\t""\t"

3. ./dataset/example_tweets.txt Temporary File To Hold Tweets For NLP POS Tagger in following format:

4. ./dataset/<train/test>ingTokenised.txt File created after running NLP POS Tagger on dataset in following format "\t""\t""\t"

5. ./dataset/final<Train/Test>ingInput.txt Combine <train/test>ingDatasetProcessed.txt <train/test>ingTokenised.txt to create dataset/final<Train/Test>ingInput.txt which contains tagged tweets with their labels in follwing format: "\t""\t""\t"

6. ./code/taskB.gs Contains actual labels for the tweets in dataset/testingTokenised.txt

7. ./code/taskB.pred Contains predicted labels for the tweets in dataset/testingTokenised.txt

Run Following Commands:

1. Remove entries from tranning dataset for which tweet is not available $python ./code/extractDataset.py ./dataset/trainingDatasetComplete.txt ./dataset/trainingDatasetProcessed.txt ./dataset/example_tweets.txt

2. POS tagging training dataset $./ark-tweet-nlp/runTagger.sh ./dataset/example_tweets.txt > ./dataset/trainingTokenised.txt

3. Combine trainingDatasetProcessed.txt trainingTokenised.txt to create dataset/finalTrainingInput.txt which contains tagged tweets with their labels $python ./code/combine.py ./dataset/trainingDatasetProcessed.txt ./dataset/trainingTokenised.txt ./dataset/finalTrainingInput.txt

4. Remove entries from testing dataset for which tweet is not available $python ./code/extractDataset.py ./dataset/testingDatasetComplete.txt ./dataset/testingDatasetProcessed.txt ./dataset/example_tweets.txt

5. POS tagging testing dataset $./ark-tweet-nlp/runTagger.sh ./dataset/example_tweets.txt > ./dataset/testingTokenised.txt

6.Combine testingDatasetProcessed.txt testingTokenised.txt to create dataset/finalTestingInput.txt which contains tagged tweets with their label $python ./code/combine.py ./dataset/testingDatasetProcessed.txt ./dataset/testingTokenised.txt ./dataset/finalTestingInput.txt

7. Train the model on ./dataset/finalTrainingInput.txt and test it on /dataset/finalTestingInput.txt, generating ./code/taskB.gs ./code/taskB.pred $python ./code/main.py ./dataset/finalTrainingInput.txt ./dataset/finalTestingInput.txt

8. Finds precison and recall $python ./code/findf1.py ./code/taskB.gs ./code/taskB.pred