Rohit

Network Congestion | Self project [ May' 20 - Jun' 20 ]

• Built a predictive model with 84% accuracy to overcome the problem of the network congestion of the Telecommunication Industry • Visualized the data completely using EDA and found reasons for particular congestion using 36 different features of 80k+ datasets • Used advanced machine learning algorithms like XGboost & statistical libraries like Scikit learn to filter out variables with high correlation

• Made new features with the help of feature engineering, correlation, and confusion matrix which increased the model accuracy by 4%