Journal of Information System Exploration and Research

Breast Cancer Diagnosis Utilizing Artificial Neural Network (ANN) Algorithm for Integrating Multi-Omics Data and Clinical Features

Rofik Rofik, Fani Artiyani; Dwika Ananda Agustina Pertiwi — Mon, 08 Jul 2024 00:00:00 +0000

Breast cancer is one of the most common diseases affecting women worldwide, with a significant impact on patient's health and quality of life. Despite advances in medical technology and research, breast cancer diagnosis remains a challenge due to its complexity involving various biological and clinical factors. Several previous studies have focused on detecting this disease with optimal accuracy, but the selection of appropriate algorithms and methods is key to achieving this goal. This study aims to improve the accuracy of breast cancer diagnosis by using the ANN algorithm and data balancing method, SMOTE. This research uses Multi-Omic data and Clinical Features obtained in general from Kaggle. The research process is carried out in several stages, namely Data Collection, Preprocessing, Oversampling, Modeling, and Evaluation. This research successfully obtained an increase in accuracy, which was able to achieve an accuracy of 99.30%. This research shows that early detection of breast cancer with ANN algorithm and data balancing using SMOTE can improve accuracy performance in early detection of breast cancer. Given the use of data in this study is not too large, it is recommended for further research to use a larger dataset to validate the strength of the model that has been built on more varied data.

The Influence of Determining the K-Value on Improving the Diabetes Classification Model using the K-NN Algorithm

Nanda Putri Korina, Budi Prasetiyo, Ade Anggian Hakim, M Rivaldi Ali Septian — Sat, 27 Jul 2024 00:00:00 +0000

Diabetes mellitus is still an important health problem globally, so it requires an efficient classification model to help determine a patient's diagnosis. This study aims to determine the K-value on the accuracy performance of the diabetes classification model using the K-Nearest Neighbors (K-NN) algorithm. This research utilizes a simulated dataset generated through interaction with ChatGPT, we investigate various K-values in the K-NN model and assess its accuracy using a confusion matrix. Based on experiments, we found that the K-NN classification model with a K=6 obtained an optimal accuracy of 97.62%. Thus, our findings highlight the important role of selecting optimal K-values in improving the performance of diabetes classification models.

Digit and Mark Recognition Using Convolutional Neural Network for Voting Digitization in Indonesia

Mandasari Mandasari, Bagus Al Qohar — Tue, 30 Jul 2024 00:00:00 +0000

Digitization of voting results in Indonesia is essential to ensure the accuracy and integrity of the election process. This research introduces an innovative approach that uses Convolutional Neural Networks (CNN) for handwritten number recognition and tally mark recognition. This research uses a dataset obtained from Kaggle. The research process is conducted in several stages, namely Data Collection, Preprocessing, Data Sharing, Modelling, and Evaluation. The results showed that the proposed model achieved an accuracy of 98%. This research shows that using the CNN algorithm for handwritten number recognition and tally mark recognition can improve accuracy and efficiency in digitizing voting results. It is expected that this research can make a significant contribution to the development of a more reliable digital voting system. Future research is recommended to use a larger dataset to validate the strength of the model, which has been built on more varied data.

Implementation of Least Significant Bit Steganography to Secure Text Messages in Images

Wiyan Herra Herviana, Djuniadi Djuniadi — Tue, 30 Jul 2024 00:00:00 +0000

Steganography is a technique of securing secret messages in other messages that are not known. Simulation of the steganography method using the Least Significant Bit technique is used to change the last bit in one byte of data by using a text message as the container medium. This study aims to implement the security of text messages in images using the Least Significant Bit technique which is supported by the steganography method. Simulation techniques are used to conduct studies using Cryptool2 which can describe the concept of cryptography. The results obtained from this study regarding the security of text message insertion into an image in *.jpg and *.png format with 5 sampling trials are (1) the encrypted image cannot be distinguished directly through human eyes, (2) there is an increase in file size the image after being encrypted with an average for five trials is 0.31%, this increase depends on the length of the text message and a key to be inserted, the longer the insertion, the larger the resulting file size, (3) The higher the resolution of the image where the description encryption is inserted, the longer the process required, (4) The simulation time of steganographic decryption is faster than steganographic encryption. The decryption simulation process is the same as 50% of the encryption process.

Inception ResNet v2 for Early Detection of Breast Cancer in Ultrasound Images

Tue, 30 Jul 2024 00:00:00 +0000

Breast cancer is one of the leading causes of death in women. Early detection through breast ultrasound images is important and can be improved using machine learning models, which are more accurate and faster than manual methods. Previous research has shown that the use of the CNN (Convolutional Neural Network) algorithm in breast cancer detection still does not achieve high accuracy. This study aims to improve the accuracy of breast cancer detection using the Inception ResNet v2 transfer learning method and data augmentation. The data is divided into training, validation and testing data consisting of 3 classes, namely Benign, Malignant and Normal. The augmentation process includes rotation, zoom, and rescale. The model trained using CNN and Inception ResNet v2 showed good performance by producing the highest accuracy of 89.72% in the training data evaluation data and getting 90% accuracy in the prediction test stage with data testing. This study shows that the combination of data augmentation and the Inception ResNet v2 architecture can improve the accuracy of breast cancer detection in CNN models.