data mining - 02/05/2018 07:48 EDT

Some FAQs regarding the project 2 are given below. 1) Can I modify at python files other than [login to view URL] No, you should not. Your changes should be limited to the [login to view URL] 2) Can I use external libraries to improve the classification F1 score? Yes, you can. Some possible changes are given as examples from the last two discussions are uploaded in the content section of the Project 2. 3) Can I just focus on the classification metrics for optimization? Yes, you should optimize the code for the highest average score F1 of the classfication metrics. 4) Can I ignore "executing" the [login to view URL] No, you cannot. because the create_model only create model for the training and and "evaluate_model" function in [login to view URL] is executed only when you run eval.py. So you should 1) optimize the code for good classification metrics and after you are happy with the classification metrics (high F1 score), 2) Run the [login to view URL] to make sure everything is working, and 3) include the ndcg shown in your report. 5) My ndcg and avg F1 score, is not changing that much, what should I do? Don't worry, I am not looking for good metrics. I want to see how much effort you put on tuning the classification metrics, grades will be based on that. 6) I have a high avg F1-score but my ndcg went down, What should I do? As I mentioned earlier, there is no final answer for this project, report the findings in your report. _________________________________________________________________________________________________ git link: [login to view URL]