Claim Missing Document
Check
Articles

Found 2 Documents
Search

Integrating Random Forest And Forward-Chaining Inference For Automated Coffee Quality Classification Using Sensory Standards sari, ika yusnita; Khairunnisa, Khairunnisa; Rahmi, Elvika; Rangkuti, Siti Rafiah; Rachmadini, Haliza Suci
The IJICS (International Journal of Informatics and Computer Science) Vol. 9 No. 3 (2025): November
Publisher : Universitas Budi Darma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30865/ijics.v9i3.9585

Abstract

The increasing consumption of coffee has driven the need for a fast and consistent coffee quality assessment process. The quality of specialty coffee is generally determined through cupping tests based on sensory attributes; however, this method still relies heavily on panelist subjectivity and requires considerable time and cost. This study aims to develop an automated system for specialty coffee quality classification by integrating the Random Forest algorithm and Forward Chaining inference logic. Random Forest is employed to perform initial classification and identify the importance level of sensory attributes, while Forward Chaining functions as a rule-based system to validate and explain the classification results. The study utilizes 207 coffee sensory profile data samples with 11 attributes based on the Specialty Coffee Association (SCA) cupping standards. The experimental results show that the Random Forest model achieves optimal performance with 100% accuracy, precision, recall, and F1-score, with Total Cup Points identified as the most dominant attribute. The integration of these two methods produces an accurate, consistent, and explainable coffee quality classification system in accordance with SCA standards.
IMPLEMENTATION OF PRINCIPAL COMPONENT ANALYSIS (PCA) IN DIMENSION REDUCTION BASED ON INDONESIAN HEALTH DATA Rangkuti, Siti Rafiah; Fadhillah, Nurul; Sari, Rita Novita; Faigle, Ulrich
Journal of Mathematics and Scientific Computing With Applications Vol. 6 No. 2 (2025)
Publisher : Pena Cendekia Insani

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.53806/jmscowa.v6i2.1313

Abstract

Indonesian health data for 2024 has multidimensional characteristics with a large number of interconnected variables, leading to high complexity in the analysis and visualization process. This complexity poses a challenge in generating information that is easy to understand and can support data-driven decision-making. This research aims to implement the Principal Component Analysis (PCA) method as a technique for dimension reduction and visualization of Indonesian health data. The research method used is a quantitative approach with descriptive-exploratory secondary data analysis. The research stages include data pre-processing, PCA implementation, principal component determination, variable contribution analysis, and data visualization using scatter plots and biplots. The research results show that PCA is able to significantly reduce the number of variables while still retaining most of the main information contained in the data. Principal component analysis-based visualization produces clearer and more easily interpretable patterns and structures in health data. Thus, PCA has proven effective in simplifying the complexity of national health data and supporting the presentation of more informative and actionable information for decision-making in the health sector.