Forecasting Cryptocurrency Markets: Predictive Modelling Using Statistical and Machine Learning Approaches
Journal Title: International Journal of Current Science Research and Review - Year 2024, Vol 7, Issue 10
Abstract
The rapidly evolving landscape of cryptocurrency markets presents unique challenges and opportunities. The significant daily variations in cryptocurrency exchange rates lead to substantial risks associated with investments in crypto assets. This study aims to forecast the prices of cryptocurrencies using advanced machine learning models. Among seven models that were tested for their prediction and validation efficiency, Neutral Networks performed the best with minimum error. Thus, Long Short-Term Memory (LSTM) neural networks were used for predicting future trends. LSTM model is well-suited for analyzing complex dependencies in financial data. Starting with historical data collection, data preprocessing, feature engineering, normalization and integrative binning, a comprehensive Exploratory Data Analysis (EDA) was conducted on 50 cryptocurrencies. Top performers were identified based on criteria such as trading volume, market capitalization, and price trends. The LSTM model was implemented using Python to predict 90-day price movements data to check intricate patterns and relationships. Model performance was validated by performance metrics such as MAE and RMSE. The findings align with the Adaptive Market Hypothesis (AMH) which suggests that cryptocurrency markets exhibit dynamic efficiency influenced by evolving market conditions and investor behavior. The study shows the potential of machine learning models in financial economics and their role in enhancing risk management strategies and investment decision-making processes.
Authors and Affiliations
Shafeeq Ur Rahaman, Patchipulusu Sudheer, Mahe Jabeen Abdul,
Keywords
Related Articles
- EP ID EP747067
- DOI 10.47191/ijcsrr/V7-i10-15
- Views 52
- Downloads 0
Parenting styles and juvenile delinquency in Zambia
This main objective of this study was to explore the relationship between maternal and paternal parenting styles and juvenile delinquency among juvenile offenders at Katombora, Nakambala and Insakwe approved schools in Z...
Teachers’ Digital Literacy Skills and Implementation of Competency Based Curriculum in Junior Schools in Westlands Sub County, Kenya
The study explored the impact of teachers’ digital literacy skills on the implementation of the Competency-Based Curriculum (CBC) in Junior Schools in Westlands Sub-County, Nairobi, Kenya. The hypothesis tested was H01:...
The Development of RBL-STEM Learning Materials to Improve Students' Information Literacy in Solving Rainbow Antimagic Coloring Problem for ETLE Technology
Students often struggle to solve complex mathematical problems in real-world contexts due to low information literacy skills. To improve information literacy skills, an effective learning approach can be RBL-STEM, which...
Maternal Characteristics and Iron Intake as a Factors of Iron Deficiency Anemia among Pregnant Women
Pregnancy is a critical period for fetal growth and development, nutritional issues can significantly affect both maternal and fetal health. This phase is marked by increased nutritional needs to support fetal developmen...
Optimizing User Retention for a Digital Recruitment App with CLM and the Hooked Model
As the digital recruitment market in Indonesia becomes increasingly competitive, it is crucial for Chatterhire, a job portal tech startup, to retain a strong user base for sustainability and growth. Retention is vital fo...