Repository containing a portfolio of my subset of data science projects completed by me(Sreekanth Settur) for academic, self learning, and hobby purposes. Presented in the form of iPython Notebooks. Below are the briefings for my projects that are added as part of portfolio.

1. Insurance claim prediction: Implemented more than 20 classification models along with performance metrics. Attached python notebook using machine learning algorithms under sklearn.

2. Customer segmentation using un-supervised learning: Implemented k-means clustering on un-labeled data to further classify the customers, based on various factors (attributes) in the data.

3. Image recognition using CNN Used pre-loaded keras dataset (CIFAR-10) to detect amount 10 labeled images.

4. MNIST digit recognition: A typical deep learning use case to work on. Used in-built keras dataset and implemented solution using fully connected dense network and also with convolution neural network.

5. Stock Price prediction: Implemented using intrinio API to collect live/historical stock price data, which is stored in mongoDB collection locally. Can be extended to export into mongoDB Atlas. Data is fetched into keras sequential model. From this point, two keras models are deployed to

-- Solve a multi-class classification problem:

• positive profit zone
• No profit zone
• Negative profit zone

-- Regression problem to estimate the approximate profit. Solution is implemented using LSTM with keras.