aakashjhawar / analyticsvidhya-india-ml-hiring-hackathon-2019

Analytics Vidhya India ML Hiring Hackathon 2019

Analytics Vidhya - India ML Hiring Hackathon 2019

My solution to AV - ML Hackathon 2019.

Loan Delinquency Prediction

Loan default prediction is one of the most critical and crucial problem faced by financial institutions and organizations as it has a noteworthy effect on the profitability of these institutions. In recent years, there is a tremendous increase in the volume of non – performing loans which results in a jeopardizing effect on the growth of these institutions. Therefore, to maintain a healthy portfolio, the banks put stringent monitoring and evaluation measures in place to ensure timely repayment of loans by borrowers. Despite these measures, a major proportion of loans become delinquent. Delinquency occurs when a borrower misses a payment against his/her loan.

Given the information like mortgage details, borrowers related details and payment details, our objective is to identify the delinquency status of loans for the next month given the delinquency status for the previous 12 months (in number of months)

Data Description

train.zip

train.zip contains train.csv. train.csv contains the training data with details on loan as described in the last section

Data Dictionary:

• loan_id Unique loan ID
• source Loan origination channel
• financial_institution Name of the bank
• interest_rate Loan interest rate
• unpaid_principal_bal Loan unpaid principal balance
• loan_term Loan term (in days)
• origination_date Loan origination date (YYYY-MM-DD)
• first_payment_date First instalment payment date
• loan_to_value Loan to value ratio
• number_of_borrowers Number of borrowers
• debt_to_income_ratio Debt-to-income ratio
• borrower_credit_score Borrower credit score
• loan_purpose Loan purpose
• insurance_percent Loan Amount percent covered by insurance
• co-borrower_credit_score Co-borrower credit score
• insurance_type 0 - Premium paid by borrower, 1 - Premium paid by Lender
• m1 to m12 Month-wise loan performance (deliquency in months)
• m13 target, loan deliquency status (0 = non deliquent, 1 = deliquent)

test.zip

test.zip contains test.csv which has details of all loans for which the participants are to submit the delinquency status - 0/1 (not probability)

sample_submission.zip

sample_submission.zip contains the submission format for the predictions against the test set. A single csv needs to be submitted as a solution.

Evaluation Metric

Submissions are evaluated on F1-Score between the predicted class and the observed target.

Approach

I first visualized the dataset to check if there is any correlation between any columns. Then I plotted various columns to check the behaviour of the dataset. After plotting the graphs, I used LabelEncoder to encode the categorical variables. Then I used Feature Scaling to rescale the dataset. Then I used XGBoost Classifier to train the dataset. I also used GridSearch to select best parameters of the model.

How to use

Install the dependencies and libraries to get started.

$ pip3 install -r requirements.txt

Start the Jupyter Server to access the notebook.

$ jupyter notebook

Or run the Python3 file.

$ python3 ml_hiring_hackathon.py

F1 Score: 0.35233

Secured Rank 5 on Public leader board. Link to Public Leaderboard

Secured Rank 23 on Private leader board. Link to Private Leaderboard

analyticsvidhya-india-ml-hiring-hackathon-2019's Issues

One hot encoding, pair wise Correlation greater than .7

Hi Aakash,
i was going through you code, i have few queries ,why you have not used One hot encoding after label encoding. 2nd question is as i saw you have not removed feature with pairwise correlation if value greater than .7 . Can you explain this please.