Yana, Baktiar Yudha (2025) ANALISIS PERBANDINGAN ALGORITMA XGBOOST DAN KNN DALAM KLASIFIKASI PENYAKIT DAUN PADI MENGGUNAKAN EKSTRAKSI HOG. Undergraduate thesis, UPN Veteran Jawa Timur.
This is the latest version of this item.
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (cover)
21081010324.-cover.pdf Download (816kB) |
|
Text (bab 1)
21081010324.-bab1.pdf Download (253kB) |
|
Text (bab 2)
21081010324.-bab2.pdf Restricted to Repository staff only until 27 November 2027. Download (586kB) |
|
Text (bab 3)
21081010324.-bab3.pdf Restricted to Repository staff only until 27 November 2027. Download (978kB) |
|
Text (bab 4)
21081010324.-bab4.pdf Restricted to Repository staff only until 27 November 2027. Download (2MB) |
|
Text (bab 5)
21081010324.-bab5.pdf Download (240kB) |
|
Text (daftar pustaka)
21081010324.-daftarpustaka.pdf Download (190kB) |
|
Text (lampiran)
21081010324.-lampiran.pdf Restricted to Repository staff only Download (228kB) |
Abstract
Rice (Oryza sativa) is a staple food source that is susceptible to disease infection, resulting in decreased productivity and quality of agricultural produce. Early identification of health problems in rice leaves plays a crucial role in maintaining stable food security. This study compared the performance of two algorithms, XGBoost and KNN, to classify rice leaf diseases using a Histogram of Oriented Gradients (HOG) feature extraction approach. The data used were divided into four disease categories: Bacterial Leaf Blight, Brown Spot, Healthy Leaf, and Leaf Blast, with a total of 180 primary and 1,400 secondary data points obtained from the Rice Leaf Disease Dataset on the Kaggle platform. Data preprocessing included resizing, augmentation, grayscale analysis, and pixel value standardization before feature extraction using the HOG method. Model evaluation was conducted with three variations of training and testing data splits: 90:10, 85:15, and 80:20. Research findings show that the KNN algorithm with parameter k = 2 at a 90:10 split is able to achieve an optimal accuracy level of 91.61% when tested on secondary data, while the XGBoost algorithm with a configuration of n_estimators 400, max_depth 4, and learning_rate 0.1 obtained an accuracy of 88.89% on primary data. The integration between the HOG method and the KNN algorithm shows the best performance with a superior level of accuracy, proving the effectiveness of this strategy in identifying diseases in rice leaves. The results of this study can be a basis for the development of a more efficient and accurate automatic detection system. Keywords: HOG extraction, image classification, KNN, rice leaf disease, XGBoost
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | T Technology > T Technology (General) | ||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Informatics | ||||||||||||
| Depositing User: | Baktiar Yudha Yana | ||||||||||||
| Date Deposited: | 28 Nov 2025 07:59 | ||||||||||||
| Last Modified: | 28 Nov 2025 07:59 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/47105 |
Available Versions of this Item
- ANALISIS PERBANDINGAN ALGORITMA XGBOOST DAN KNN DALAM KLASIFIKASI PENYAKIT DAUN PADI MENGGUNAKAN EKSTRAKSI HOG. (deposited 28 Nov 2025 07:59) [Currently Displayed]
Actions (login required)
 |
View Item |