Al'Adzkiya International of Computer Science and Information Technology Journal
Vol 3, No 2 (2022)

Analysis K-Nearest Neighbors (KNN) in Identifying Tuberculosis Disease (Tb) By Utilizing Hog Feature Extraction

Muhathir, Muhathir (Unknown)
Sibarani, Theofil Tri Saputra (Unknown)
Al-Khowarizmi, Al-Khowarizmi (Unknown)



Article Info

Publish Date
01 Dec 2022

Abstract

Pulmonary tuberculosis is an infectious disease caused by Microbacterium tuberculosis, which is one of the lower respiratory tract disease, which is largely in the pulmonary tissue of the lung infection and then undergoes a process known as the primary focus of Ghon. Because the disease is difficult and takes a long time to decide the patient is affected by the disease Tuberkolusis, then the detection of the patient affects Tuberkolusis by utilizing the K-NN method as a classification and HOG as feature extraction. Results of the classification of positive diagnosis with a total of 234 samples from 330 samples or successfully recognizable Sebasar 70.90%, while the classification result is a negative diagnosis with the amount of 240 samples from 330 samples or successfully identified by 72.72%. The results of the study showed the image classification of the X-ray Set Tuberculosis using the method K-NN and HOG feature with cross-validation 5 folds with 71.81% accuracy. Keyword : tuberculosis, K-NN, HOG.

Copyrights © 2022






Journal Info

Abbrev

AIoCSIT

Publisher

Subject

Computer Science & IT

Description

Computer Science, Computer Engineering and Informatics: Data Science Artificial Intelligence, Machine Learning, Neural Network, Computer Architecture, Parallel and Distributed Computer, Pervasive Computing, Computer Network, Embedded System, Human—Computer Interaction, Virtual/Augmented Reality, ...