**Learning from CQA Data**
						**Team No : 15**				

1. Enviornment : Python version : 3.6.1 Anaconda version : 4.4.0 Modules required to run the program: a. numpy b. scikit c. nltk d. pandas e. matplotlib f. re g. statistics h. math

2. File Descriptions: APCsvgenerator.py : CSV generator for answer quality prediction
APFeatures.py : Class to store features for answers quality prediction
APHelperClass.py : Helper class for answer quality prediction
APQuestionAnswer.py : For answer score prediction converts posts to questions and answers using post type id i.e question(id:1) or answer(id:2)
Badges.py : To store data from badges.xml file
CQAAnalysis.py : Main script to run the program.(Entry point)
Comments.py : To store data from comments.xml file
Csvgenerator.py : To generate CSV data to predict long term value of questions
Features.py : To store features for long term value
Files.py : To map the xml files.
GraphPlotting.py : To plot the accuracy graphs for long term value prediction
HelperClass.py : It contains helper functions for long term value prediction
LRClassifier.py : Logistic Regression for QA score prediction and long term value prediction.
LongTermValue.py : Entry point for long term value prediction called by CQAAnalysis.py
MLPClassifier.py : Multi Layer Perceptron for QA score prediction.
MLStripper.py : Strips html tags and code fragments.
NBClassifier.py : Naive Bayes Classifier for QA score prediction and long term value prediction.
PostHistory.py : To store data from posthistory.xml file
PostLinks.py : To store data from postlinks.xml file
Posts.py : To store data from posts.xml file
PredictQuestionAnswerQuality.py : Entry point for Question Answer score prediction called by CQAAnalysis.py
QPCsvgenerator.py : CSV generator for question quality prediction
QPFeatures.py : To store features for question quality prediction
QPHelperClass.py : Helper class for question quality prediction
QPQuestionAnswer.py : For Question score prediction converts posts to questions and answers using post type id i.e question(id:1) or answer(id:2)
QuestionAnswer.py : For Long term value prediction, converts posts to questions and answers using post type id i.e question(id:1) or answer(id:2)
QuestionTagger.py : Tag prediction model
SVMClassifier.py : Support Vector Machines Classifier for QA score prediction and long term value prediction.
ScoreClassifier.py : Used for question answer score prediction using multiple models.
Tags.py : To store data from tags.xml file
Users.py : To store data from users.xml file
ViewCountClassifier.py : Used for long term value prediction using multiple models.
Votes.py : To store data from votes.xml file


3. Using the program :

1. Unzip CQA_Analysis.zip.
   
2. Change directory to CQA_Analysis
   
3. Make sure these folders exist.
   a. Data/ai.stackexchange.com
   b. CSV Files
   c. CSV Files_A
   d. CSV Files_Q
   e. Images_LTV
   f. Images_A
   g. Images_Q
   
4. $python3 CQAAnalysis.py
   

**4. Expected output: ** Following the above instructions the program will train the data and test and create the following accuracy graphs

**5. Results: ** The results are in the form of graphs generated from the analysis which are added in the images folder.
1. Images\Logistic RegressionClassifier.png : Accuracy vs time for Long term value prediction using Logistic Regression
2. Images\Support vector machinesClassifier.png : Accuracy vs time for Long term value prediction using SVM.
3. Images\Navie BayesClassifier.png : Accuracy vs time for Long term value prediction using Naive Bayes
4. Images_Q\Accuracy.PNG: Performance of differnt models for Question score prediction
5. Images_Q\MLP_Classifier.png :Accuracy vs k value of k-fold for Question score prediction using MLP
6. Images_Q\Gaussian_Naive_Bayes.png : Accuracy vs k value of k-fold for Question score prediction using GNB
7. Images_Q\Support_Vector_Machine.png : Accuracy vs k value of k-fold for Question score prediction using SVM
8. Images_Q\Logistic_Regression.png : Accuracy vs k value of k-fold for Question score prediction using LR
9. Images_Q\coefficient.PNG : Coefficents for each features in LR for Question score prediction
10. Images_A\Accuracy.PNG: Performance of differnt models for Answer score prediction
11. Images_A\MLP_Classifier.png :Accuracy vs k value of k-fold for Answer score prediction using MLP
12. Images_A\Gaussian_Naive_Bayes.png : Accuracy vs k value of k-fold for Answer score prediction using GNB
13. Images_A\Support_Vector_Machine.png : Accuracy vs k value of k-fold for Answer score prediction using SVM
14. Images_A\Logistic_Regression.png : Accuracy vs k value of k-fold for Answer score prediction using LR
15. Images_A\coefficient.PNG : Coefficents for each features in LR for Answer score prediction