Abstract
Cannabis is the most used drug around the world with the highest risks and associated criminal problems in many countries. This research describes the process of classifying online posts to identify cannabis use problems and their associated risks as early as possible. We annotated 11,008 online posts, which we used to build robust classification models. We tested classical and deep learning classifiers. Different CNN- and RNN-based models proved to be promising approaches to detect cannabis use posts. Our system can be used by authorities (such as parents or doctors) to monitor cannabis use-related posts. It could raise an alarm to the relevant authorities to take necessary interventions to analyze the cannabis use risks associated with the posts. To the best of our knowledge, this is the first study that uses deep learning methods successfully to detect cannabis use from any text or online posts. We tested our deep learning models on the SubUse-Cann unseen dataset which contains 17,099 tweets. It is an imbalanced dataset with only 6.9% positive cannabis use tweets. Our CNN-based model performed the best with an accuracy of 95.25% and an F1-score of 92.83% for classifying cannabis use.
Supported by the Natural Sciences and Engineering Research Council of Canada (NSERC).
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
UNODC World Drug Report 2020 https://www.unodc.org/unodc/press/releases/2020/June/media-advisory---global-launch-of-the-2020-world-drug-report.html.
- 2.
Scikit-learn 1.0.1 2021: https://scikit-learn.org/stable/supervised_learning.html.
References
Bahdanau, D., Cho, K., Bengio, Y.: Neural machine translation by jointly learning to align and translate. arXiv preprint arXiv:1409.0473 (2014)
Bansal, D., Chhikara, R., Khanna, K., Gupta, P.: Comparative analysis of various machine learning algorithms for detecting dementia. Procedia Comput. Sci. 132, 1497–1502 (2018).. https://doi.org/10.1016/j.procs.2018.05.102
Bergman, B.G., Dumas, T.M., Maxwell-Smith, M.A., Davis, J.P.: Instagram participation and substance use among emerging adults: the potential perils of peer belonging. Cyberpsychol., Behav. Soc. Netw. 21(12), 753–760 (2018)
Çöltekin, Ç., Rama, T.: Drug-use identification from tweets with word and character n-grams. In: Proceedings of the EMNLP Workshop SMM4H: The 3rd Social Media Mining for Health Applications Workshop & Shared Task, pp. 52–53 (2018)
Cox, M.J., Janssen, T., Gabrielli, J., Jackson, K.M.: Profiles of parenting in the digital age: associations with adolescent alcohol and marijuana use. J. Stud. Alcohol Drugs 82(4), 460–469 (2021)
Derczynski, L., et al.: Analysis of named entity recognition and linking for tweets. Inf. Process. Manag. 51(2), 32–49 (2015)
Fischer, B., Russell, C., Sabioni, P., Van Den Brink, W., Le Foll, B., Hall, W., Rehm, J., Room, R.: Lower-risk cannabis use guidelines: a comprehensive update of evidence and recommendations. Am. j. Public Health 107(8), e1–e12 (2017)
George, M.J., Ehrenreich, S.E., Burnell, K., Kurup, A., Vollet, J.W., Underwood, M.K.: Emerging adults’ public and private discussions of substance use on social media. Emerg. Adulthood 9(4), 408–414 (2021)
Hassanpour, S., Tomita, N., DeLise, T., Crosier, B., Marsch, L.A.: Identifying substance use risk based on deep neural networks and instagram social media data. Neuropsychopharmacology 44(3), 487–494 (2019)
Hu, H., et al.: An insight analysis and detection of drug-abuse risk behavior on Twitter with self-taught deep learning. Comput. Soc. Netw. 6(1), 1–19 (2019)
Hu, H., et al.: An ensemble deep learning model for drug abuse detection in sparse twitter-sphere. In: MedInfo, pp. 163–167 (2019)
Jenhani, F., Gouider, M.S., Said, L.B.: Lexicon-based system for drug abuse entity extraction from twitter. In: BDAS, pp. 692–703 (2016)
Kalchbrenner, N., Grefenstette, E., Blunsom, P.: A convolutional neural network for modelling sentences. arXiv preprint arXiv:1404.2188 (2014)
Kingma, D.P., Ba, J.: Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014)
Koratana, A., Dredze, M., Chisolm, M.S., Johnson, M.W., Paul, M.J.: Studying anonymous health issues and substance use on college campuses with YIK yak. In: AAAI Workshop: WWW and Population Health Intelligence (2016)
Mahata, D., Friedrichs, J., Shah, R.R., et al.: # phramacovigilance-exploring deep learning techniques for identifying mentions of medication intake from twitter. arXiv preprint arXiv:1805.06375 (2018)
Manning, C.D., Surdeanu, M., Bauer, J., Finkel, J.R., Bethard, S., McClosky, D.: The stanford corenlp natural language processing toolkit. In: Proceedings of 52nd annual meeting of the association for computational linguistics: system demonstrations, pp. 55–60 (2014)
Menon, A., Farmer, F., Whalen, T., Hua, B., Najib, K., Gerber, M.: Automatic identification of alcohol-related promotions on twitter and prediction of promotion spread. In: 2014 Systems and Information Engineering Design Symposium (SIEDS), pp. 233–238. IEEE (2014)
Orabi, A.H., Buddhitha, P., Orabi, M.H., Inkpen, D.: Deep learning for depression detection of twitter users. In: Proceedings of the Fifth Workshop on Computational Linguistics and Clinical Psychology: From Keyboard to Clinic, pp. 88–97 (2018)
Pedregosa, F., et al.: Scikit-learn: machine learning in python. J. Mach. Learn. Res. 12, 2825–2830 (2011)
Pennebaker, J.W., Chung, C.K.: Expressive writing: connections to physical and mental health. In: Friedman, H.S. (ed.) The Oxford Handbook Of Health Psychology, pp. 417–437. Oxford University Press (2011)
Pennington, J., Socher, R., Manning, C.: Glove: global vectors for word representation. In: Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp. 1532–1543. Association for Computational Linguistics, Doha, Qatar (October2014). https://doi.org/10.3115/v1/D14-1162,https://www.aclweb.org/anthology/D14-1162
Raja, B.S., Ali, A., Ahmed, M., Khan, A., Malik, A.P.: Semantics enabled role based sentiment analysis for drug abuse on social media: a framework. In: 2016 IEEE Symposium on Computer Applications & Industrial Electronics (ISCAIE), pp. 206–211. IEEE (2016)
Sarker, A., O’Connor, K., Ginn, R., Scotch, M., Smith, K., Malone, D., Gonzalez, G.: Social media mining for toxicovigilance: automatic monitoring of prescription medication abuse from twitter. Drug Safety 39(3), 231–240 (2016)
Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.: Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Rese. 15(1), 1929–1958 (2014)
Vázquez, A.L., et al.: Innovative identification of substance use predictors: machine learning in a national sample of Mexican children. Prevent. Sci. 21(2), 171–181 (2020)
Yadav, S., et al.: “When they say weed causes depression, but it’s your fav antidepressant’’: knowledge-aware attention framework for relationship extraction. PLoS one 16(3), e0248299 (2021)
Acknowledgment
This research was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC), the Ontario Centres of Excellence (OCE), and SafeToNet.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Ibrahim, D., Inkpen, D., Osman, H.A. (2023). Identifying Cannabis Use Risk Through Social Media Based on Deep Learning Methods. In: Rutkowski, L., Scherer, R., Korytkowski, M., Pedrycz, W., Tadeusiewicz, R., Zurada, J.M. (eds) Artificial Intelligence and Soft Computing. ICAISC 2022. Lecture Notes in Computer Science(), vol 13589. Springer, Cham. https://doi.org/10.1007/978-3-031-23480-4_9
Download citation
DOI: https://doi.org/10.1007/978-3-031-23480-4_9
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-23479-8
Online ISBN: 978-3-031-23480-4
eBook Packages: Computer ScienceComputer Science (R0)