This is the python notebook for the quora question pair similarity on Kaggle: Kaggle competition website.
The problem gives us a training dataset of question pair and the testset. The task is to analyze and categorize each pair of questions so that we can determine whether the two questions are similar or not.
The training dataset and the test dataset can befound using Kaggle dataset :Kaggle Quora question pair dataset.
In addition, I use word2vec pre-train vector from Glove to support the sentiment analysis for my model: Glove word2vec
A slide show to show how well my prediction works including the accuracy, precision and recall value. It also shows some of the algorithm that I use for this particular problem
Question pair similarityA final technical report on the quora question pair problem, as well as detail discussion of the algorithm and technique that I use for the final prediction.