Search

This study aims to extract feature values from thermal breast images using Gabor Filter feature extraction, focusing on mean, variance, kurtosis, skewness, and entropy, and to evaluate the performance of three classification methods, namely K-Nearest Neighbor (KNN), Support Vector Machine (SVM), and Artificial Neural Network (ANN). Breast cancer is a serious health issue, especially for women, as it can potentially lead to death. In an effort to reduce mortality risks, research is conducted to detect cancer early, including the use of thermography. This method utilizes temperature from objects to detect cancer, where different temperature patterns in breast areas affected by cancer can be observed due to increased blood flow. The study employs thermal images from the Database for Mastology Research (DMR), consisting of 150 images, with 108 healthy and 42 diseased images. Texture features are extracted using Gabor Filter with variations in scale and orientation angles. The results are tested using several classification methods, with ANN showing the highest accuracy of 88.88%, followed by KNN with 86.66% and SVM with 84.44%. These findings confirm that thermography, along with texture feature extraction and machine learning algorithms, can effectively detect breast cancer early, offering potential for early diagnosis and effective disease management. [ABSTRACT FROM AUTHOR]

As time goes by, the occurrence of online crimes has significantly increased. Therefore, there is a need to improve the handling of online crimes. This is in line with the Kapolri Priority Program, which is to make the Indonesian National Police (Polri) human resources excel in the era of Police 4.0. This study was conducted to identify the types of crimes under the ITE Law or online crimes, evaluate the results of the K-NN (K-Nearest Neighbor) prediction method on online crimes, and analyze the profile of online crime victims based on the dataset. The research method used in this study was quantitative, supported by qualitative analysis. The results showed that the majority of online crimes that occurred in the jurisdiction of the Metro Bekasi Police Resort were online fraud, hacking/cracking, and online insult/defamation. Meanwhile, the K-NN method prediction results showed the best accuracy of 98,67% with a training data ratio of 90% and 10% testing data, using k = 1, 3, 5, 9. The profile of online crime victims revealed that the majority were males, with most victims being professionals, and the age of the victims mostly occurred in early adulthood. Most victims lived in urban areas, and the crime scene often occurred in the city centre. The time of occurrence was often in the morning and early hours of the day, with Wednesdays having a high incidence of crimes. [ABSTRACT FROM AUTHOR]