This repo holds the code for my MA thesis: "Using Machine Learning and Google Searches to Predict Stock Market Changes".
The main script that I used to generate features, run random forest and SVM models, and plot data.
Packages Used in this script:
- ggplot2
- tidyverse
- fpp2
- TTR
- pracma
- data.tables
- caret
- mlbench
- randomForest
- gridExtra
- ggpubr
- mlbench
- e1071
A Python script I created and used to collect and normalize daily Google search data. Special thanks and credit goes to a module called pytrends
, a pseudo API created by GeneralMills for grabbing data from Google Trends. See the module here.
Modules used in this script:
- pytrends
- datetime
- pandas
- matplotlib
- time
- random