This was another Kaggle competition in association with Reva University and Swiss Re. In this competition competitors was expected to build a classifier which can predict whether a person has Autism or not.
In this challenge first I performed some exploratory data analysis to derive insights from it regarding how different factors affect whether a person has autism or not. Data provided in this competition was not a standard one so, I have performed Data Cleaning as well before performing the data Analysis. Data Imbalance was also one of the challenge which was supposed to be handled by the competitors.
I used LogisticRegression, Support Vector Machine and XGBoostClassifier from the sklearn library but the performance achieved by LogisticRegression was 85% along with the validation accuracy of 82% which implies that the model is not overfitting the data was able to generalize well on the training data.
By using this model's predictions I scored 88% accuracy on the test data and secured 9th position out of 120 competitors in this competition.
Checkout my Jupyter Notebook on kaggle for a better readable interface.
Thankyou for reading.