Detailed descriptions can be found in the Methodology section.

Below is a brief overview of our methodology, spanning from the initial data collection through to strategy backtesting:

• Data Collection: We collect, organize and reduce over 25 years of historical data price and financial statement data for publicly traded companies.
• Data Preprocessing: The historical data undergoes preprocessing, utilizing a variety of machine-learning techniques to manage missing data points and outliers. Additionally, the dataset underwent rigorous manual quality control to further enhance its reliability and accuracy for model training.
• AI Model Training: We manage our dataset and conduct both preprocessing and model training operations using the Google Vertex AI platform. It is important to note that our AI Model is dynamic as we continue to feed new data into the model from time to time, all with the aim of continuously updating and improving the model.
• Data Analysis: Subsequently, the trained models were deployed to assign labels to stock at various historical points, based on the stocks financial metrics to create the ProPicks Strategies. The ProPicks Strategies are updated periodically, based on the updated ratings generated by the AI model analysis.
• Backtesting: The ProPicks Strategies were then backtested using our proprietary backtesting module, in order to assess how such strategies would have performed historically.