Solution for India_Hacks-Machine Learning

Solution which placed 15th on LeaderBoard.

https://www.hackerearth.com/machine-learning-india-hacks-2016/machine-learning/will-bill-solve-it/

Data Sets:

Both training and testing dataset consist of 3 files :-

1) User File:

With Attributes of a User:

user_id - the user id
skills - all his skills separated by the delimiter '|'
solved_count - number of problems solved by the user
attempts - total number of incorrect submissions done by the user
user_type : type of user (S - Student, W - Working, NA - No Information Available)

2) Problem File:

Attribute related to a Problem :

problem_id - the id of the problem
level - difficulty of the problem (Very-Easy, Easy, Easy-Medium, Medium, Medium-Hard, Hard)
accuracy - the accuracy score for the problem
solved_count - number of people who have solved it
error_count - number of people who have solved it incorrectly
rating - star (quality) rating of the problem on scale of 0-5
tag1 - tag of the problem representing the type e.g. Data Structures
tag2 - tag of the problem
tag3 - tag of the problem
tag4 - tag of the problem
tag5 - tag of the problem

3) Submissions File:

Problem User interaction and final results for each attempt a user made to a solve a particular problem.

user_id - the id of the user who made a submission
problem_id - the id of the problem that was attempted
solved_status - indicates whether the submission was correct (SO : Solved or Correct solution, AT : Attempted or Incorrect solution )
result - result of the code execution (PAC: Partially Accepted, AC : Accepted, TLE : Time limit exceeded, CE : Compilation Error, RE : Runtime Error, WA : Wrong Answer)
language_used - the lang used by user to code the solution
execution_time - the execution time of the solution

Approach

count - A custom feature. Weighs are assigned to user's skills and based on the skills that a user has, the value is calculated.

Used Ensemble of RandomForestClassifier, ExtraTreesClassifier and GradientBoostingClassifier.

Installations Required#####

python sklearn >=0.16 numpy pandas

Instructions to run the code

python model.py