The topic is very relevant and comprehensively explained. The growing importance of machine learning techniques in finance supports the relevance of this topic, as these techniques can enhance market forecasting and decision-making processes. The aim and objectives are well-explained and clearly defined. The thesis is well-structured, with a logical flow throughout all sections, enhancing the readability and comprehensibility of the thesis. The analytical approach of the thesis is robust, employing various techniques to optimize LSTM model parameters and enhance predictive accuracy. The solutions to the research objectives are appropriate. The candidate, however, forgot to explain that the Random Walk theory is a part of market efficiency, which is why it supports the RW theory. The thesis employs appropriate data preprocessing techniques, ensuring data quality and suitability for the chosen forecasting methods. The selection of research tools and methods is well justified, and the list of references is relevant and comprehensive. However, the thesis could benefit from including several more references, discussing them in the study, and linking them to the findings. Ensa has contributed significantly to selecting and justifying the research model, developing the methodology, and comprehensively addressing the research objectives. The thesis results are well-justified and interpreted, offering practical guidance for investors and policymakers. The layout fulfills the requirements of the Regulations for master thesis preparation and defense. However, the overall presentation could be improved. Such studies are rich in exhibits showing different trends, but not bar exhibits. Additionally, I recommended that the candidate exclude codes from the main body of the text and present them in the Appendix; however, the codes remain in the main text. Figure 7 should be a table. Figure 2 is excellent and explains the entire work. ADF results should be presented in table format (Figure 8). The paper does not contain any elements of plagiarism. Originality is 96.8%.