Abstract:
This paper Examines effects on the health of the population of various drugs in Bangladesh through machine learning Techniques. I analyze the impact of different types of drugs on Bangladeshi society and investigate the predictors of drug dependence among individuals with physical and mental health issues. Data were collected from various sources, like surveys and interviews.Took authorize data from some of well-known reputed psychiatrist. More than five hundred (500+) data respondents in the Bangladeshi were analyzed using different machine learning algorithms. The results showed that physical health was the strongest predictor of drug dependence, followed by mental health, financial problems, reasons, and age. The study also revealed that the majority of drug users were aged between 18 to 25 years. I found the second number of drug user’s age between 26-30 after analyzing the data. Where most of the user’s age are 20-25 now. These findings suggest that physical and mental health issues are important risk factors for drug dependence and should be considered when developing strategies to reduce the prevalence of drug use and also it has important implications for drug prevention and treatment programs in Bangladesh. The research further highlighted the importance of preventive measures to reduce the overall consumption of drugs in Bangladesh. According to the report, machine learning algorithms can be used to create more effective preventative measures and gain a better understanding of the effects of various drug types in Bangladesh. Furthermore, the article makes the case that these tactics ought to concentrate on lowering the use of drugs like marijuana, yaba, heroin, or alcohol and encouraging the use of healthier substitutes. I predicted seven distinct medications using five different algorithms, and my accuracy ranged from 45% to 65%. Next, I applied PCA and achieved more than 98%. Accuracy. Basically, I got the result most efficient and less errorless. The highest accuracy got on Phensedy l00%,using KNN.
