The paper develops an algorithm for making long-term (up to three months ahead) predictions of volatility reversals based on long memory properties of financial time series. The approach for computing fractal dimension using sequence of the minimal covers with decreasing scale is used to decompose volatility into two dynamic components: specific and structural. We introduce two separate models for both, based on different principles and capable of catching long uptrends in volatility. To test statistical significance of its abilities we introduce several estimators of conditional and unconditional probabilities of reversals in observed and predicted dynamic components of volatility. Our results could be used for forecasting points of market transition to an unstable state.
Class description: introductory course on developing trading algorithms and algotrading industry as a whole. Class is held in financial lab with 9 Bloomberg terminals; students are learning to fast-prototype algorithms using R language and real data from markets. Focus is on practice, but good understanding of underlying theory is a must. Knowledge of Bloomberg platform and R language is an advantage.
Prerequisites: basics of financial markets, technical and fundamental analysis, financial mathematics, modern portfolio theory, statistics and probability theory.
Modern portfolio theory assumes that decisions are made by individual agents. In reality most investors are involved in group decision-making. The paper renders group decision-making process by means of random forest algorithm, which could significantly improve prediction of weak learners by combining them into one model with superior performance. We combine technical, fundamental and sentiment analysis in order to generate views on different asset classes. Then the portfolio model is built using copula opinion-pooling under views generated by random forest. The model is backtested and results are compared with the ones obtained using traditional asset allocation techniques.