The dataset was taken from here and the data description can be found here. Various attributes were analysed and the effect of nicotine usage on them was predicted as well. Different methodologies implemented were one attribute classifier, KNN algorithm with K value as 1 & 3 and compared with Fisher Linear Discriminant method. The results obtained were interesting where it was found out that few attributes were able to predict the effect significantly than the 1NN algorithm. But, the Fisher’s linear Discriminant method stood out of other algorithms with high accuracy and very low value of error.

Ultimately, it was found out that psychological predisposition to nicotine usage does exist, which means that usage of nicotine regularly does affect our psychology and broadly the 7 classifiers used in the research. Further studies can be carried out to find out the actual impact of it and the intensity of impact such that the governments can advertise the issues accordingly in order to create awareness among the people.