Work Log

GOAL

Extract the hidden information from reviews and analyze the most important issue for restaurant

Stage 1: Tag Train Data

status:	done
problem:

1. lose location tag

Stage 2: Train data to get model

status:	done
problem:

Stage 3: Classify Selected Restaurant Reviews

status:	doing
problem:

1. Add another classifier like SVM

Stage 4: Extract Keyword from Each Class

status:	undone
problem:

Stage 5: Separate Positive and Negative Keyword

status:	undone
problem:

Unsolved Issue

1. Deprecate the other class
2. Stage 4
3. Stage 5
4. MultiTag instead of single classifier
5. Log the appropriate parameter (NaiveBayes)
6. Improvement of feature extraction
7. Assign the part to each group member
8. Study SVM classifier api
9. Related User Activeness and Reliability to Selected Reviews
10. Trainig data via iterating diffience sizes of training datasets
11. Verify the correctness of created model based on Naive Bayes Model