Fundamentální a technická analýza vývoje cen softs komodit

In this master thesis we elaborate in detail on models for forecasting prices of agriculture crops for futures market. The commodities are as follows: coffee, cocoa, sugar and cotton. For our models we acquired fundamental weather data and data from COT report, which only had short history. On the other hand, we worked with data containing technical indicators (EMA, RSI, Momentum, Bollinger bands, MACD a Stochastic), which had long history. ARIMAX, VAR and VECM models were developed in fundamental analysis. Vector auto-regression model contained data with short history. For technical analysis we used regression tree and random forest algorithms, which evaluated data sets with different length of history. Subsequently models for given models profitability calculation were developed. Profitability was evaluated using achieved t – statistics and calculated p – value from cumulative distribution function of t – distribution. At the end of this thesis we compare profitability of models and we multiple-test significant models.