JISKA (Jurnal Informatika Sunan Kalijaga)

Identifikasi Kematangan Buah Pisang Berdasarkan Variasi Jarak Menggunakan Metode K-Nearest Neighbor

2024-01-31T13:55:47+07:00

This research aims to identify the level of ripeness of kepok bananas based on the color of their skin using the K-Nearest Neighbor (K-NN) method. Bananas are an important commodity in Indonesia, and various ripeness levels need to be identified. The current process of identifying banana ripeness is still done manually, which requires a lot of labor and tends to be subjective. The K-NN method is used to classify bananas based on their skin color. This research involves the collection of banana images with three ripeness levels (raw, ripe, and overripe) and the extraction of RGB color features from these images. Three distance methods, namely Euclidean, Minkowski, and Manhattan, are also employed to compare accuracy results. The evaluation results of this research show that the accuracy value for the Euclidean distance method is 84%, the Minkowski distance method is 82%, and the Manhattan distance method is 80%. Thus, the findings indicate that the K-NN method and the Euclidean distance method provide good results in identifying the ripeness level of bananas. By implementing the K-NN algorithm, this research attempts to address the weaknesses of the time-consuming and subjective manual identification process, with the hope of providing a more accurate and efficient solution for the banana industry. The results of this research can be used to automate the identification process of banana ripeness levels and improve efficiency in banana sorting. It is expected that this research can provide practical benefits to the community and serve as a basis for further research in this field.

Segmentasi Pelanggan E-Commerce Menggunakan Fitur Recency, Frequency, Monetary (RFM) dan Algoritma Klasterisasi K-Means

2024-02-16T16:28:55+07:00

The rapid growth in the e-commerce industry demands the development of smarter and more focused marketing strategies. One approach that can be applied is customer segmentation using various features such as Recency, Frequency, and Monetary (RFM), along with machine learning-based clustering methods. The objective of this study is to design and develop a web-based e-commerce customer segmentation application using a combination of RFM features and clustering methods. The study proposes the K-Means algorithm and compares it with K-Medoids and Fuzzy C Means using publicly available e-commerce datasets. Experimental results showed that the K-Means algorithm outperformed K-Medoids and Fuzzy C Means (FCM) based on the Silhouette Score of 0.67305, Davies Bouldin Index of 0.51435, and Calinski Harabasz Index of 5647.89. Through analysis and testing, the designed application has proven effective in grouping customers into relevant segments. These segments are divided into three categories: Loyal, Need Attention, and Promising, visualized in a web-based application dashboard using Streamlit. The developed application allows e-commerce business owners and users from the business, management, and marketing divisions to categorize customers based on transaction data. The results of this study are expected to provide valuable insights to e-commerce management and marketing professionals who are facing increasingly fierce competition.

Analisis Performa Normalisasi Data untuk Klasifikasi K-Nearest Neighbor pada Dataset Penyakit

2024-03-23T12:44:49+07:00

This study investigates four normalization methods (Min-Max, Z-Score, Decimal Scaling, MaxAbs) across prostate, kidney, and heart disease datasets for K-Nearest Neighbor (K-NN) classification. Imbalanced feature scales can hinder K-NN performance, making normalization crucial. Results show that Decimal Scaling achieves 90.00% accuracy in prostate cancer, while Min-Max and Z-Score yield 97.50% in kidney disease. MaxAbs performs well with 96.25% accuracy in kidney disease. In heart disease, Min-Max and MaxAbs attain accuracies of 82.93% and 81.95%, respectively. These findings suggest Decimal Scaling suits datasets with few instances, limited features, and normal distribution. Min-Max and MaxAbs are better for datasets with numerous instances and non-normal distribution. Z-Score fits datasets with a wide range of feature numbers and near-normal distribution. This study aids in selecting the appropriate normalization method based on dataset characteristics to enhance K-NN classification accuracy in disease diagnosis. The experiments involve datasets with different attributes, continuous and categorical data, and binary classification. Data conditions such as the number of instances, the number of features, and data distribution affect the performance of normalization and classification.

Implementasi Data Augmentation untuk Klasifikasi Sampah Organik dan Non Organik Menggunakan Inception-V3

2024-03-30T10:22:32+07:00

The surge in global waste, particularly in Indonesia, with a total of 36.218 million tons per year, has become an urgent issue. Challenges in waste management are increasingly complex due to the lack of public understanding and awareness in classifying types of waste. One systemic approach to address waste classification issues involves the use of machine learning technology to categorize waste into two main types: organic and non-organic. The data used in this study comes from a Kaggle website dataset comprising 25,500 entries. This research employs a transfer learning approach with the Inception-V3 architecture and data augmentation implementation. Transfer learning is chosen for its proven performance in image data classification, while data augmentation is implemented to introduce diversity to the dataset. The research stages include business understanding, data preprocessing, data augmentation, data modelling, and evaluation. The results show that the use of transfer learning with the Inception-V3 approach and data augmentation implementation achieves an accuracy rate of 94%, which falls into the excellent category.

Implementasi K-Means Clustering pada Pengelompokan Pasien Penyakit Jantung

2024-05-22T15:32:46+07:00

Heart disease is a prominent global health concern, necessitating early identification and patient grouping for effective management. This study employs the K-Means clustering algorithm with a medical dataset of 303 patients, encompassing various attributes. These include Age, Gender, Chest Pain Type, Blood Pressure, Serum Cholesterol Level, Fasting Blood Sugar, Resting Electrocardiographic Results, Maximum Heart Rate, Angina, ST Depression, and Slope of the ST Segment. The goal is to categorize patients into four clusters based on chest pain types, a crucial symptom indicating disease severity. The computation concludes after the sixth iteration, revealing Cluster 1 (27 patients), Cluster 2 (135 patients), Cluster 3 (15 patients), and Cluster 4 (126 patients). Collaborative analysis with medical experts highlights that Cluster 1, mainly comprising older males, exhibits high-risk indicators. While this grouping aids in personalized treatment strategy development, further clinical validation involving more experts and datasets is imperative for enhanced reliability.

Pelabelan Sentimen Berbasis Semi-Supervised Learning menggunakan Algoritma LSTM dan GRU

2024-05-22T15:46:34+07:00

In the sentiment analysis research process, there are problems when still using manual labeling methods by humans (expert annotation), which are related to subjectivity, long time, and expensive costs. Another way is to use computer assistance (machine annotator). However, the use of machine annotators also has the research problem of not being able to detect sarcastic sentences. Thus, the researcher proposed a sentiment labeling method using Semi-Supervised Learning. Semi-supervised learning is a labeling method that combines human labeling techniques (expert annotation) and machine labeling (machine annotation). This research uses machine annotators in the form of Deep Learning algorithms, namely the Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) algorithms. The word weighting method used in this research is Word2Vec Continuous Bag of Word (CBoW). The results showed that the GRU algorithm tends to have a better accuracy rate than the LSTM algorithm. The average accuracy of the training results of the LSTM and GRU algorithm models is 0.904 and 0.913. In contrast, the average accuracy of labeling by LSTM and GRU is 0.569 and 0.592, respectively.

Integrating Retrieval-Augmented Generation with Large Language Model Mistral 7b for Indonesian Medical Herb

2024-07-22T16:02:04+07:00

Large Language Models (LLMs) are advanced artificial intelligence systems that use deep learning, particularly transformer architectures, to process and generate text. One such model, Mistral 7b, featuring 7 billion parameters, is optimized for high performance and efficiency in natural language processing tasks. It outperforms similar models, such as LLaMa2 7b and LLaMa 1, across various benchmarks, especially in reasoning, mathematics, and coding. LLMs have also demonstrated significant advancements in addressing medical queries. This research leverages Indonesia’s rich biodiversity, which includes approximately 9,600 medicinal plant species out of the 30,000 known species. The study is motivated by the observation that LLMs, like ChatGPT and Gemini, often rely on internet data of uncertain validity and frequently provide generic answers without mentioning specific herbal plants found in Indonesia. To address this, the dataset for pre-training the model is derived from academic journals focusing on Indonesian medicinal herbal plants. The research process involves collecting these journals, preprocessing them using Langchain, embedding models with sentence transformers, and employing Faiss CPU for efficient searching and similarity matching. Subsequently, the Retrieval-Augmented Generation (RAG) process is applied to Mistral 7b, allowing it to provide accurate, dataset-driven responses to user queries. The model's performance is evaluated using both human evaluation and ROUGE metrics, which assess recall, precision, F1 measure, and METEOR scores. The results show that the RAG Mistral 7b model achieved a METEOR score of 0.22%, outperforming the LLaMa2 7b model, which scored 0.14%.