Data from: Why do youths initiate to smoke? A data mining analysis on tobacco advertising, peer, and family factors for Indonesian youths
Data files
Oct 16, 2024 version files 20.97 KB
-
README.md
588 B
-
research_data_tobacco_mining.csv
20.38 KB
Abstract
Global Youth Tobacco Survey (GYTS), Indonesia showed that 60,9 % of students noticed cigarette advertisements or promotions in outdoor media. Our study aimed to understand the impact of outdoor tobacco advertising and peer and family association with Youth's smoking behavior.
This study deployed a cross-sectional approach to explore factors related to youth smoking behavior, such as peers, family, and tobacco advertising. The GYTS questionnaire was adapted as the instrument and distributed to 400 students from 20 high schools to observe smoking behavior. The chosen schools based on the previous study whose classify school in hot-spot and non-hot-spot areas. This study applied a data mining approach with a decision tree to generate the models.
This study generates a decision tree model that describes the peer factor as the key to introducing Youth to smoking. The model also reveals that youth in the non-hotspot advertising area are not likely to develop Youth to smoke. The model has a performance classification of 77.5 % This study found that youth with smoking fathers are more likely to start smoking earlier, youth whose both parents are smokers, and mothers who are smokers have a confidence level of 100 % to smoke. Further research is warranted to investigate rural districts to explore any regional and socioeconomic variations.
This study deployed a cross-sectional quantitative study to explore factors related to youth smoking behaviour. This study surveyed and interviewed 400 students as a sample of high schools to observe smoking behaviour using adaptations of the GYTS questionnaires. This study applied a data mining approach with a decision tree method to generate the models. Later, the models were compared between several weighting criteria: Information Gain, Gini Index, Gain Ratio, and Accuracy. The comparison explored the most factors contributing to developing children smokers among peers, family, and tobacco advertising.
There are 400 cases in this dataset.
The attributes:
- Outdoor advertising (hotspot,non hotspot);
- Parent Smoker (No one, Father, Mother, Both of them, Do not know);
- Peer Smoker (No one, Few, Most of them, all smoker);
- Children smoker (No, Yes);
- Start to smoke ( never smoke, < 7 years old, 8-9 years old, 10-11 years old, 12-13 years old, 14-15 years old)