Vytváření investičních strategií pomocí analýzy časových řad

This bachelor thesis is focused on summary analysis of time series for prediction returns of stock titles and indexes using time series prediction methods. And, at the same time, testing whether market predictions make sense in relation to longer prediction on time series. In the prediction I use simple methods together with bootstrap and ARIMA. On these methods, I predict stock returns and indices to predict possible future developments over a five-year period. The methods I teach on data from 2009 to 2016. I then compare the actual revenue development for 2017 with the methods mentioned above and choose the best prediction method based on the measure of prediction accuracy of a forecasting method. I represent a small investor who creates a portfolio from selected stocks and indexes, which then optimizes using the Markowitz method. I model portfolio optimization according to two assumptions. The first is the assumption of a minimum standard deviation that represents risk. The second prerequisite is the best value of the sharpe ratio, which represents the best ratio between return and risk in the portfolio. Then, based on the best prediction method chosen, I adjust the portfolio and analyse how this method will affect the future composition of the portfolio.