Využití analýzy sentimentu při predikci vývoje vybraných akciových titulů

Abstract

The aim of this diploma thesis is to reconstruct the CNN Fear & Greed Index for the period from 1998 to 2026 and to assess whether the resulting sentiment indicator can be used in the construction of trading strategies. The reconstruction is based on seven components corresponding to the methodology of the original index. Two versions of the index are created, one with equal weights and the other with weights estimated using the OLS method. The quality of the reconstruction is assessed by comparison with historical CNN Fear & Greed Index values for the period from 2011 to 2026. Subsequently, six trading strategies are tested, with their parameters determined using the period from 1998 to 2015 and evaluated in the subsequent period from 2016 to 2026. The results indicate that the reconstructed index captures the development of the original CNN index well. None of the tested strategies outperformed the Buy and Hold benchmark based on the S&P 500 Total Return Index in terms of total return in the subsequent test period. Nevertheless, some strategies achieved a better risk-return profile, especially those based solely on long positions.