akhilsharmaa / Quora-Question-Pairs-Kaggle

Kaggle Competition, Achived 74% accuracy.

Quora Question Pairs Competition

Method 1: Random Forest

Accuracy using Random-Forest : 0.73

Apply random forest on the dataset directly, Let's see what is the accuracy.

Method 1.2: Random Forest after feature engineering

Accuracy after feature eng. using Random-Forest : 0.73

In the feature engineering: we will add this features in the data_frame

• question1_lenght - char lenght of question1
• question2_lenght - char lenght of question2
• q1_word_count - no. of words in the question1
• q2_word_count - no. of words in the question2
• word_commmon - common word in q1 & q2
• word_total - Total word in q1 + q2
• word_share - word_common / word-total