financial-statements-analysis

• The goal of this project is to design a software system which can predict a company's Return on Equity(ROE) score for the upcoming quarter quarter/year.

• Return on equity (ROE) is a measure of financial performance calculated by dividing net income by shareholders' equity. Because shareholders' equity is equal to a company’s assets minus its debt, ROE could be thought of as the return on net assets.

Design

Data preparation

Training

Features used

• PreviousRoe
• PercentOfPositiveWords
• PercentOfNegativeWords
• PercentOfLitiguousWords
• PercentOfConstrainingWords
• PercentOfStrongModalWords
• PercentOfWeakModalWords
• PercentOfUncertaintyWords
• (Label) PercentChangeInRoe

Master data

• Ticker
• CompanyName
• CompanyCategory
• Date
• TtmNetIncome
• ShareholdersEquity

Sample data

Run locally

• Data Preparation part is written in Java. It is a mvn project.

mvn clean install

• Run MainApplication.java

• Training is written in python.

• Run Training app to train a new model.

• Run Prediction app to start serving predictions.

Baseline comparison

• Baseline comparison for MSFT, GOOGL, ORCL & TWTR available here.

References

• Common Annual Algorithm
• loughran mcdonald financial dictionary
• CBOW based Financial Keyword Expansion
• Regression techniques