Abstract:
Pakistan has around 24-million tobacco users and annually more than one-lac people died
due to tobacco related diseases. After adoption of PSEP&PNSH-ordinance in 2002, the
government continuously tried to make strategies to reduce the tobacco consumption and
prevalence. The prime objective of this study is to evaluate the effectiveness of anti tobacco legislation by gathering the data related to tobacco consumption and adverse
effects of tobacco on human health. The dataset of the study comprises of sixteen variables,
out of which four are consider as response variables which are directly associated with
effectiveness of anti-tobacco legislation in Pakistan. We are summarizing all anti-tobacco
legislations implemented in Pakistan from 1958 to onwards and used different time series
and machine learning models to analyze the response variables and forecast their future
values up to the year-2030. The evaluation of the models assessed through MSE, MAE and
MAPE and with the help of plots of testing verses predicting data. Overall, the performance
of ML-models like linear regression & Random Forests much better as compared to other
and when trend presence in the data set, the Holt’s method of exponential smoothing
performed well. Analysis showed that, over the period anti-tobacco legislations played a
significant role to reduce the tobacco consumption and their adverse effects.
