AXOVISION develops a model of interrelated artificial intelligences which analyzes the stock market. All relevant data for share prices, from financial ratios to sentiment analysis, are being utilized. The stock portfolio generated by our prototype shows a performance of 33% p.a. and thereby outperforms the broader stock market significantly. The portfolio will be put on the market as an investment fund.
The advantage over a human is that the Machine Learning model can detect relevant and neglect irrelevant data. It is able to solve complex, interdependent coherences in short order without the interference and influence of emotions and biases. Additionally, the usage of one model is scalable at only minimal cost – unlike the services of fund managers.
Modern day financial markets are fast-paced and dynamic. This requires flexible models that constantly detect changes and adapt accordingly. Our model incorporates strategies to handle concept drifts and adjust to shifts in market behavior and ensures suitability at all times.
The unique way of integrating different machine learning approaches into a higher-level ensemble enables more course-relevant information to be processed.
Concept Drift Handling enables a constant prediction quality over time and an automatic adaptation to the market and investor behaviour.
Diverse data analysis on traditional and alternative data enables higher ROI opportunities by improving decision making in the investment process.
The implemented long/short trading strategy enables high return opportunities, unaffected by the general market performance or investor behaviour.
