Feeding a supervised machine learning model with financial indicators to predict a hold, sell or buy classification compared to the S&P 500 as a benchmark.
- Codes were aligned with the chapters in the summary (ex.: for chapter 2, see notebook document "2. Data Cleaning.ipynb")
- Data drawn from kaggle and other ressources were placed in the file "raw_data"
- Our cleaned data set can be located in the file "cleaned_data"
The task consists of creating an investment recommendation model, which is able to predict an annual stock return based on financial indicators from the US Stock market for the period of 2014 to 2018. Our model should be able to indicate the return into a buy/hold/sell classification, which will be compared with the S&P 500, a stock market index which includes the largest companies listed in the United States. Before delving into our data set, we decided to coordinate our approach so we will be able to cover each of the questions asked step by step. In our initial meeting we quickly realized which milestones we would be facing during this project. First of all, we noticed a lack of inconsistency in our data set that we will further elaborate in our first part in our summary. In our second step we outlined which algorithm will be the most suitable to apply in this project. Here we would measure the performance for each technique used to make a comparison. We are going to discuss which performance metric was applied in the corresponding section. Third, we further refine our most accurate model under the conditions such as varying the number of given features as well as adding new features. Last but not least, we would conclude our project with highlighting our achievements but also our limitations.