Quora is a website with over 300 million monthly users, where they can ask or answer questions. It is a place where people can share or gain knowledge about anything they are curious about. With such a large community, similar questions are going to be asked which may cause users to spend more time to find their answers. Which leads us to our goal in “Quora Question Pairs Classification.” Using a data set of sample Quora questions, we aim to present a model capable of determining whether a pair of questions are duplicates of each other. This can be extremely beneficial to users, allowing their questions to be answered much more quickly as well as reducing load on persons providing answers. This problem offers an opportunity to explore feature extraction, classification and deep learning techniques.
-
unzip training data into the same directory as the .py files. The dataset is in .zip format because of its large size.
-
Refer to imports for each file to ensure that all python libraries are installed.
-
To run glove embddings, use 'glove.840B.300d.txt' found at https://github.com/stanfordnlp/GloVe