${\color{lightblue} Internet-Service-Churn-Prediction}$

${\color{lightblue}Techniques \space Used}$

• Internet Service Churn Prediction using Deep Learning (Tensorflow and Keras).
• The Problem was of Classification type where we have to predict whether a customer will end the services from te given firm on the basis of certain parameters.

${\color{lightblue} Table}$

${\color{red} Serial \space No. }$	${\color{red} Work}$	${\color{red} Description }$	${\color{red} Progress }$
1.	${\color{lightgreen} Data \space Preprocessing}$	Pre-processing of Dataset downloaded from Kaggle. This included the dropping of Null values of rows, evaluation and conversion of non-integer values to integer values via Pipelining Integration.	Completed
2.	${\color{lightgreen} Data \space Analysis}$	Analysis of Dataset and evaluating general trends in the values. Encoding the Categorical data columns into One-Hot Encoding values (integer data type). Visualization done by Seaborn, Matplotlib and Plotly.	Completed
3.	${\color{lightgreen} Feature \space Extraction}$	Since the values were all integers and one-hot encoded, the only thing left to be done was feature scaling. As, none of the values were negative and there were a few outliers, so I preferred to use Min-Max Scaler.	Completed
4.	${\color{lightgreen} Splitting \space The \space Dataset}$	The Dataset is splitted into train size (75%) and test size(25%) by the help of scikit-learn python library.	Completed
5.	${\color{lightgreen} Deep \space Learning}$	Deep Learning and Model Training using the Artificial Neural Networks (ANN). There were two hidden layers present in the Neural Network. Plotted the Confusion Matrix and the Classification Report with the help of Tensorflow and Keras as a Wrapper Library.	Completed

${\color{lightblue} Made \space By}$

${\color{lightgreen} Vishu \space Kalier}$