AI for Investment Management
- Gather dataset on entities and their fundamentals over 10 year historical records
- Gather a live stream of sentiment data, including from last 10 years, via social media, news, economic indicators, and other sources
- Identify the key features for extraction
- Develop an investment strategy as preference styles (value, growth, momentum, etc)
- Develop a risk tolerance threshold with a defined decision making rulebook
- Develop a matrix for uncertainty reasoning
- Develop a target market for investment segments (retail, institutional, etc)
- Develop a value proposition
- Develop a Hybrid AI technique using a combination of probabilistic and structured approaches (e.g. deep learning, knowledge graph, generative ai, natural language processing, machine learning, natural computation, causal reasoning, gametheory, timeseries forecasting, multiagents)
- Develop a set of models via train/validate/test/tune/evaluate
- Integrate the models into a portfolio construction, trade execution, portfolio rebalancing, risk management, asset allocation, security selection, performance monitoring processes
- Link the investment strategy against core set of themed goals: grow wealth, preserve wealth, generate income
- Continuously monitor performance, retrain, and apply adaptive feedback
- Incrementally add components for explainability, ethics, regulatory compliance, assurance, due diligence checks, adverse media checks, KYC checks, PEP checks, sanctions checks, vessel embargo checks, AML checks, data breach checks, entity/industry rundown, security and privacy
- Run a regular set of policy and audit trails through the pipeline processes for transparency and governance
