Application of Machine Learning algorithms for AIX Index movements predictions - MASTER’S DISSERTATION - May 2024

Sep 17, 2026

Uploaded by Noah

The master’s dissertation investigates how advanced machine learning techniques can forecast the indices of the Astana International Exchange (AIX), a highly impactful Central Asian stock exchange. Forecast accuracy is positioned as crucial for investors, fund managers, and policy makers to make informed decisions and understand market fluctuations. The study evaluates multiple models for projecting AIX index trend performance and assesses implications for investment strategies and broader market movements. Historical AIX index share price data for 2023 is used for learning and validation, and results indicate ARIMA can be a useful instrument for forecasting AIX price dynamics.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages61

Reading Time2 h 34 min

AccessFree to download

Introduction

Literature Review

Data Collection and Preprocessing

Model Selection and Training

Linear Regression

Time Series Analysis

Autoregressive Integrated Moving Average (ARIMA)

Support Vector Machines (SVM)

Conclusion

References

Maqsut Narikbayev University International School of Economics MASTER’S DISSERTATION «Application of Machine Learning algorithms for AIX Index movements predictions» 7M04124- «Finance» May 2024 Written by: Yerzhigit Ichshanov Supervisor: Yerzhan Tokbolat Astana, 2024 Abstract There have been many efforts to predict stock market trends by numerous scholars across a wide range of areas utilizing different methods and applications. This thesis examines in detail how advanced machine learning techniques can be used to predict the indices of the Astana International Exchange (AIX) . As a relatively new but highly impactful stock exchange in Central Asia, the AIX is important in the regional financial landscape. The accurate forecasting of AIX index trends is critical for stakeholders such as investors, fund managers, and policy makers, as it allows them to take informed decisions and obtain better understanding of market fluctuations. The paper evaluates the performance of different machine learning models in projecting the trend performance of AIX indices and investigates the consequences of these predictions for investment strategies and wider market movements. For the learning and validation of machine learning models, the research uses historical data on AIX index share prices for the period of 2023 year. The study shows that all used models may give sufficient outcomes in forecasting the dynamic of AIX index. The findings demonstrate that in comparison with other models, ARIMA can become useful instrument for forecasting price dynamic on AIX and can be used by investors and traders to take well-grounded conclusions in their investment strategies. Key words: Machine Learning, AIX Index, Financial Prediction , ARIMA Model Table of Contents Introduction ..................................................................................................................... 7 Literature Review ............................................................................................................ 9 Data Collection and Preprocessing ............................................................................... 12 Model Selection and Training ........................................................................................ 13 Linear Regression .................................................................................................... 13 Time Series Analysis ............................................................................................... 17 Autoregressive Integrated Moving Average (ARIMA)…………………………...…20 Support Vector Machines (SVM) .............................................................................. 24 Data Processing……………………………………………………………………………….27 Linear Regression Results……………………………………………………………………30 Multiple Regression Results…………………………………………………………………35 ARIMA Results……………………………………………………………………………….. 38 AR Results……………………………………………………………………………………..42 Comparisons between different methods’ results………………………………………….50 Conclusion…………………………………………………………………………………...…53 References…………………………………………………………………………………......57 List of tables Table 1. Comparisons of the average prices of shares of CCBN…………………….….41 Table 2. Comparisons of the average prices of shares of KAP………………………….42 Table 3. Comparisons of the average prices of shares of KMG………………………....42 Table 4. Comparisons of the average prices of shares of HSBK………………...….......43 List of illustrations Figure 1. Ordinary share prices for the March in 2024…………………………………….24 Figure 2. Ordinary share prices of “KazMunayGas” JSC in March in 2024………………25 Figure 3. Ordinary share prices of JSC “KazAtomProm" in March in 2024………………25 Figure 4. Common share prices of JSC “Halyk Savings Bank of Kazakhstan…………..26 Figure 5. The main code of linear regression…………………………………………........27 Figure 6. Linear Regression results for CCBN…………………………………………......28 Figure 7. Linear Regressio

FAQs

What is the main objective of the dissertation?

To examine how advanced machine learning techniques can be used to predict AIX index movements and trends in the Astana International Exchange.

Which data period is used to train and validate the machine learning models?

Historical AIX index share price data for the period of 2023 is used for learning and validation.

Which model is highlighted as especially useful for forecasting AIX price dynamics?

The findings indicate that ARIMA can become useful for forecasting AIX price dynamics compared with other models.

Sep 17, 2026

Uploaded by Noah