PREDICTION OF BRONCHITIS DISEASE INDICATIONS USING CATBOOST ALGORITHM AND SHAPLEY ADDITIVE EXPLANATIONS WITH A RULE-BASED SYSTEM FOR DRUG RECOMMENDATIONS

Hakim, Lukman (2026) PREDICTION OF BRONCHITIS DISEASE INDICATIONS USING CATBOOST ALGORITHM AND SHAPLEY ADDITIVE EXPLANATIONS WITH A RULE-BASED SYSTEM FOR DRUG RECOMMENDATIONS. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text 21081010118_Covers.pdf Download (600kB)
	Text (Bab 1) 21081010118_BAB 1.pdf Download (178kB)
	Text (Bab 2) 21081010118_BAB 2.pdf Restricted to Registered users only until 23 July 2029. Download (552kB) \| Request a copy
	Text (Bab 3) 21081010118_BAB 3.pdf Restricted to Registered users only until 23 July 2029. Download (857kB) \| Request a copy
	Text (Bab 4) 21081010118_BAB 4.pdf Restricted to Registered users only until 23 July 2029. Download (5MB) \| Request a copy
	Text (Bab 5) 21081010118_BAB 5.pdf Download (120kB)
	Text (Daftar Pustaka) 21081010118_DAFTAR PUSTAKA.pdf Download (154kB)
	Text (Lampiran) 21081010118_Lampiran.pdf Download (239kB)

Abstract

Bronchitis is one of the most common respiratory diseases that requires accurate and timely medical treatment. However, the process of diagnosis and medication recommendation often relies heavily on manual analysis by medical personnel. Therefore, this study aims to develop a predictive model for bronchitis indications using the CatBoost algorithm and implement a rule-based drug recommendation system to support medical decision-making. The research methodology follows the stages of machine learning implementation, including data collection from patient symptoms at Apotek RH Farma Surabaya January 2025 period until April 2025, data preprocessing, feature engineering, labeling, hyperparameter tuning, and model evaluation using Accuracy, Precision, Recall, and F1-Score metrics. The model was tested using three training-to-testing data ratios (60:40, 70:30, and 80:20). The results show that the model with a 80:20 ratio achieved the best performance, with Accuracy 83%, Precision 75%, Recall 84%, dan F1-Score 79%. The SHAP (Shapley Additive Explanations) analysis revealed that symptoms such as coughing period and wheezing contributed the most to bronchitis prediction. Furthermore, the rule-based recommendation system successfully mapped patient symptoms to relevant Over-the-Counter (OTC) drug lists and provided consultation suggestions when bronchitis indications were detected. Based on these findings, the integration of the CatBoost prediction model and rule-based system proved effective in improving diagnostic accuracy and the relevance of drug recommendations.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Via, Yisti Yulia	NIDN198604252021212001	yistivia.if@upnjatim.ac.id
Thesis advisor	Puspaningrum, EvaYulia	NIDN1980705202112002	evapuspaningrum.if@upnjatim.ac.id

Subjects:

T Technology > T Technology (General) > T58.6-58.62 Management Information Systems

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Mr. Lukman Hakim

Date Deposited:

23 Jul 2026 03:32

Last Modified:

23 Jul 2026 03:32

URI:

https://repository.upnjatim.ac.id/id/eprint/57291