Final_Project_sp_500

K-Means Clustering:

https://github.com/thedankline/Final_Project_sp_500/blob/master/DC%20and%20ML.ipynb

Regression:

https://github.com/thedankline/Final_Project_sp_500/blob/master/machine%20learning%20regression.ipynb

Web Scraping:

https://github.com/thedankline/Final_Project_sp_500/blob/master/Final%20project%20first%20scratch%20code%2C%20data%20scraping.ipynb

https://github.com/thedankline/Final_Project_sp_500/blob/master/Final%20Project%20second%20scratch%20code%2C%20data%20scraping.ipynb

This project uses machine learning K-means clustering and support vector regression to analyze S&P 500 stock data to help predict which stocks to invest in. The data was collected from webscraping data from Yahoo Finance using BeautifulSoup4 in Jupyter Notebook. The data did not come out perfect so it needed to be filtered. The data was then saved to a CSV file. The data in the CSV file was used to make a scatter plot of the stock's P/E Ratio vs Dividend Yields. Using K-means clustering the scatter plot was split into five clusters. The first cluster (cluster 0) is the cluster of stocks most likely to gain value based on our analysis.

The data collection for the support vector regression model was different. The data was collected using IEXcloud API calls. The stocks in the first cluster from the K-means clusering were used in the support vector regression. Three types of support vector regressions where used to fit the data; Linear, Polynomial, and RBF (Radial Basis Function). The model was fitted using the stock's open price and the the date. This was all done in Jupyter Notebook.

The packages used in this projects are listed below, they can all be installed using a pip install.

bs4

requests

pprint

pandas

numpy

pylab

scipy

math

sklearn

matplotlib

copy