Elucidating the Impact of Opium Consumption on Blood Parameters Using Network Analysis and Machine Learning

Despite its prevalent use, the specific biochemical effects of opium on blood parameters havenot been comprehensively elucidated. Utilizing data from the comprehensive Persian FASA cohort [۱],our study harnesses the power of advanced statistical modeling and machine learning techniques tosystematically investigate these effects. By integrating traditional substance use research with cuttingedgebioinformatics, this study aims to shed light on the intricate biochemical dynamics induced byopium consumption, thereby filling a critical knowledge gap in the field. The study involved ۱۰,۱۳۸participants. It started with thorough data preprocessing, along with the implementation of strictinclusion criteria. This process guaranteed that individuals with pre-existing medical conditions werenot included in the study. In the analysis based on machine learning, ۸۰% of the data was used in thetraining phase and ۲۰% for testing. We then embarked on a detailed correlation analysis of all availableparameters. Parameters with a correlation coefficient exceeding |۰.۳| were selected for further networkanalysis. This analysis was conducted using Cytoscape software, leading to the construction of adetailed correlation network, distinctly highlighting the relationships between various bloodparameters. Our results revealed unique correlations in the opium user group, such as RBC-MCHC,PLT-MPV, TG-HDL.C, and CHOL-GGT, which were absent in the normal group. To quantify thedifferences between healthy and opium-addicted individuals, we employed the random forest algorithm,achieving an accuracy of ۸۲%, precision of ۹۲%, recall of ۶۹%, and an F۱ score of ۷۹%. The metricsdemonstrate the model's effectiveness and highlight the physiological effects of opium on bloodparameters. This study highlights the efficacy of combining mathematical approach to unravel thecomplex effects of opium on blood parameters [۲]. The random forest algorithm was used to identifyunique biochemical patterns in opium users, highlighting the potential of advanced computationaltechniques in biomedical research.

کلیدواژه ها:

Opium ، Machine learning ، Network analysis ، Blood