Identifying Tobacco Control Policy Drivers: A Neural Network Approach

This paper presents a neural network approach to investigating Australian smokers' quit motivations that affect their quit attempts. Based on the data from the International Tobacco Control Four Country Survey, Neural network (NN) models are developed to identify smokers' quitting motivations as smokers' quit motivations are significant factors in predicting smokers' quit attempts. In order to identify the underlying tobacco control policies from these quitting motivations, principle component analysis is used to group individual attributes into 4 tobacco control policy drivers: Personal Concerns, Price, Social Restrictions and Social Encouragement which are related to specific tobacco control policies. To examine the impact of these tobacco control policy drivers on smokers' making a quit attempt, a set of NN models using 4 policy drivers are also built. Experimental results indicate that in comparison with cigarette price and social restrictions, educating smokers of health benefits from quitting and social encouragement for cessation of smoking have more impacts in encouraging smokers to make a quit attempt.