Wihayati
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Wihayati has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Ferry Wahyu Wibowo
Universitas AMIKOM Yogyakarta
10 shared publications Akhmad Dahlan
Universitas Amikom Yogyakarta,Informatics Management Department,Yogyakarta,Indonesia
1 shared publication- Andi Sunyoto
Universitas AMIKOM Yogyakarta
1 shared publication Bayu Setiaji
Universitas Amikom Yogyakarta,Department of Informatics,Yogyakarta,Indonesia
1 shared publicationHindriyanto Dwi Purnomo
Department of Information Technology, Satya Wacana Christian University, Salatiga, Indonesia
1 shared publication- Hindriyanto Dwi Purnomo
Satya Wacana Christian University
1 shared publication
Latest Publications
- 8 July 2026
- Published by Springer Nature in Journal of Astrophysics and Astronomy
- Vol. 47 (2) , 37
Forecasting horizon Dst index based on solar wind data using deep learning and generative artificial intelligence
- 8 July 2026
- Published by Springer Nature in Journal of Astrophysics and Astronomy
- Vol. 47 (2) , 37
Forecasting horizon Dst index based on solar wind data using deep learning and generative artificial intelligence
- 26 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 373-379
Bisecann: A Deep Learning Model from the Bisection Mechanism and Neural Network for Time Series Cases
Modeling in deep learning (DL) makes many people curious about a solvable case. This approach serves to discover regularities/patterns and relationships within the framework of large data sets with the lowest possible error rate. The development of this technique uses programming to learn from the past or part of the object, and then this information is used to identify data. The vanishing gradient is one of the weaknesses of DL; one way to overcome this is by adding a bias factor to the artificial neural network (ANN) model to shift the activation function to the right or left. The bias factor is usually constant; some modeling uses old data values. This paper aims to create a new deep learning model architecture by adopting the mechanism of one of the closed numerical methods, namely Bisection. The actual function is to replace the flexible bias value in the ANN and depends on the current and previous data input. The use of these two methods in this paper is then called BISECANN. This paper also presents accuracy observations to determine the level of model-learning evaluation in several cases. It compares several evaluation metrics from other DL methods in predicting the same case.
- 26 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 373-379
Bisecann: A Deep Learning Model from the Bisection Mechanism and Neural Network for Time Series Cases
Modeling in deep learning (DL) makes many people curious about a solvable case. This approach serves to discover regularities/patterns and relationships within the framework of large data sets with the lowest possible error rate. The development of this technique uses programming to learn from the past or part of the object, and then this information is used to identify data. The vanishing gradient is one of the weaknesses of DL; one way to overcome this is by adding a bias factor to the artificial neural network (ANN) model to shift the activation function to the right or left. The bias factor is usually constant; some modeling uses old data values. This paper aims to create a new deep learning model architecture by adopting the mechanism of one of the closed numerical methods, namely Bisection. The actual function is to replace the flexible bias value in the ANN and depends on the current and previous data input. The use of these two methods in this paper is then called BISECANN. This paper also presents accuracy observations to determine the level of model-learning evaluation in several cases. It compares several evaluation metrics from other DL methods in predicting the same case.
- 12 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 557-562
Classification of Wireless Sensor Networks and Internet of Things Attacks Based on Chimpanzee Leader Election Optimization - Convolutional Neural Networks
The vulnerability of Wireless Sensor Networks (WSN) and the Internet of Things (IoT) to cyber attacks such as Blackhole, Flooding, Grayhole, and Time Division Multiple Access (TDMA) scheduling drives the need for accurate and resource-efficient detection solutions as one of the contributions in edge computing security. This study proposes a hybrid attack classification system by combining the Chimpanzee Leader Election Optimization (CLEO) algorithm for optimal feature selection on Convolutional Neural Networks (CNN) for attack pattern extraction. Inspired by the chimpanzee leadership hierarchy, CLEO successfully determines the optimal values of the number of neurons and the learning rate on CNN with reduced and modified datasets. CNN classifies attacks by analyzing temporal-spatial patterns of network traffic. The experimental results show that the accuracy for the training data set is 100 %, and the accuracy for the testing dataset is 98 %. This combination of CLEO-CNN not only overcomes the limitations of conventional methods in balancing accuracy and efficiency, but also offers a practical solution for real-time security on critical WSN and IoT infrastructures.
- 12 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 557-562
Classification of Wireless Sensor Networks and Internet of Things Attacks Based on Chimpanzee Leader Election Optimization - Convolutional Neural Networks
The vulnerability of Wireless Sensor Networks (WSN) and the Internet of Things (IoT) to cyber attacks such as Blackhole, Flooding, Grayhole, and Time Division Multiple Access (TDMA) scheduling drives the need for accurate and resource-efficient detection solutions as one of the contributions in edge computing security. This study proposes a hybrid attack classification system by combining the Chimpanzee Leader Election Optimization (CLEO) algorithm for optimal feature selection on Convolutional Neural Networks (CNN) for attack pattern extraction. Inspired by the chimpanzee leadership hierarchy, CLEO successfully determines the optimal values of the number of neurons and the learning rate on CNN with reduced and modified datasets. CNN classifies attacks by analyzing temporal-spatial patterns of network traffic. The experimental results show that the accuracy for the training data set is 100 %, and the accuracy for the testing dataset is 98 %. This combination of CLEO-CNN not only overcomes the limitations of conventional methods in balancing accuracy and efficiency, but also offers a practical solution for real-time security on critical WSN and IoT infrastructures.
- 20 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 304-309
Analysis of Round-Robin and Random Scheduling Algorithms on Sensor Node Activation Towards Latency of Wireless Sensor Network
A Wireless Sensor Network (WSN) is a network system consisting of wireless sensor nodes that collect and transmit data. One crucial factor that affects WSN performance is latency, which is the time required to send data to the gateway. High latency can cause data delivery delays, resulting in data becoming irrelevant. At the same time, receiving real-time data quickly is the primary monitoring method. In addition, high latency can cause greater energy consumption because the sensor node device must be active longer. One way to overcome latency is to implement traffic management techniques such as scheduling data delivery or using priority delivery methods, which can reduce congestion and latency. This paper discusses the effect of latency on WSN using scheduling algorithms, namely Round-Robin and Random scheduling. The results were then taken 30 times to obtain a statistical comparison. The results obtained in the experiment received a Round-Robin scheduling algorithm that was superior to the random scheduling algorithm with a mean and standard deviation value of 12.68 and 2.33, respectively.
- 20 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 304-309
Analysis of Round-Robin and Random Scheduling Algorithms on Sensor Node Activation Towards Latency of Wireless Sensor Network
A Wireless Sensor Network (WSN) is a network system consisting of wireless sensor nodes that collect and transmit data. One crucial factor that affects WSN performance is latency, which is the time required to send data to the gateway. High latency can cause data delivery delays, resulting in data becoming irrelevant. At the same time, receiving real-time data quickly is the primary monitoring method. In addition, high latency can cause greater energy consumption because the sensor node device must be active longer. One way to overcome latency is to implement traffic management techniques such as scheduling data delivery or using priority delivery methods, which can reduce congestion and latency. This paper discusses the effect of latency on WSN using scheduling algorithms, namely Round-Robin and Random scheduling. The results were then taken 30 times to obtain a statistical comparison. The results obtained in the experiment received a Round-Robin scheduling algorithm that was superior to the random scheduling algorithm with a mean and standard deviation value of 12.68 and 2.33, respectively.
- 11 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 409-414
Performances of Chimpanzee Leader Election Optimization and K-Means in Multilevel Color Image Segmentation
Image segmentation is one of the essential approaches in image processing and analysis required for object detection and background separation. This paper aims to analyze the performance and capabilities of the Chimpanzee Leader Election Optimization (CLEO) algorithm and the K-means clustering technique. The CLEO algorithm is an optimization model used to optimize the segmentation of color objects in images. At the same time, the objective function in finding the best values utilizes the K-means method. Determining the number of clusters in this paper uses the elbow method, where this method functions to obtain the optimal number of clusters in multilevel color image segmentation. This paper displays random images taken from a download library based on flower, animal, and transportation categories. The results of the elbow method on the three selected images in this image segmentation research have taken the number of K values in these categories, namely 4, 5, and 6, respectively. It was to demonstrate the hybrid model capabilities of the CLEO and K-means. The method for evaluating the results of this hybrid model in segmenting images used the Intersection over Union (IoU) metric, Jaccard Index, Pixel Identity, and Localization Error. The IoU metric results for these categories got a score above 99%, the Jaccard index all got a score of 100%, and Pixel Identity got a score below 1.4 %. In contrast, Localization Error scores in the categories used have yielded 15.9%, 29.3%, and 7.7%, respectively.
- 11 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 409-414
Performances of Chimpanzee Leader Election Optimization and K-Means in Multilevel Color Image Segmentation
Image segmentation is one of the essential approaches in image processing and analysis required for object detection and background separation. This paper aims to analyze the performance and capabilities of the Chimpanzee Leader Election Optimization (CLEO) algorithm and the K-means clustering technique. The CLEO algorithm is an optimization model used to optimize the segmentation of color objects in images. At the same time, the objective function in finding the best values utilizes the K-means method. Determining the number of clusters in this paper uses the elbow method, where this method functions to obtain the optimal number of clusters in multilevel color image segmentation. This paper displays random images taken from a download library based on flower, animal, and transportation categories. The results of the elbow method on the three selected images in this image segmentation research have taken the number of K values in these categories, namely 4, 5, and 6, respectively. It was to demonstrate the hybrid model capabilities of the CLEO and K-means. The method for evaluating the results of this hybrid model in segmenting images used the Intersection over Union (IoU) metric, Jaccard Index, Pixel Identity, and Localization Error. The IoU metric results for these categories got a score above 99%, the Jaccard index all got a score of 100%, and Pixel Identity got a score below 1.4 %. In contrast, Localization Error scores in the categories used have yielded 15.9%, 29.3%, and 7.7%, respectively.
- 24 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 324-329
Chimpanzee Leader Election Optimization - Support Vector Machine Classifier-Based Analysis of Handwritten Digits Recognition
Hybrid models in machine learning (ML) for a case usually have good tendencies if they get the right conditions. This paper aims to classify the recognition of handwritten digits using hybrid modeling between the chimpanzee leader election optimization (CLEO) algorithm and support vector machine (SVM) classifiers. Handwritten digit recognition is a form of pattern recognition. Pattern recognition in images refers to the segmentation of detected image objects. The various patterns of handwritten digits will be related to the results and selection of patterns used to identify and equalize the model's perception of a particular pattern. The handwritten digits in this paper consist of the numbers 0 to 9. The amount of data is 1384 data. These data have then been divided into training and testing data of 984 and 400, respectively. Because the handwritten digit dataset is unbalanced, the data needs to be balanced using Synthetic Minority Over-Sampling Technique (SMOTE). The CLEO algorithm finds the best parameter values for SVM to operate in classification. It indicates the central role of the CLEO algorithm in optimizing the SVM model so that it tends to have the best evaluation results. Assigning an appropriate model in solving a case is an absolute necessity so that it can identify the data pattern and predictions on new data can be made. The hybrid model between the CLEO algorithm and the SVM classifier in this paper has obtained accuracy results for recognizing handwritten digits of 93%. This model was then used to compare between digits. However, the results were entirely satisfactory.
- 24 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 324-329
Chimpanzee Leader Election Optimization - Support Vector Machine Classifier-Based Analysis of Handwritten Digits Recognition
Hybrid models in machine learning (ML) for a case usually have good tendencies if they get the right conditions. This paper aims to classify the recognition of handwritten digits using hybrid modeling between the chimpanzee leader election optimization (CLEO) algorithm and support vector machine (SVM) classifiers. Handwritten digit recognition is a form of pattern recognition. Pattern recognition in images refers to the segmentation of detected image objects. The various patterns of handwritten digits will be related to the results and selection of patterns used to identify and equalize the model's perception of a particular pattern. The handwritten digits in this paper consist of the numbers 0 to 9. The amount of data is 1384 data. These data have then been divided into training and testing data of 984 and 400, respectively. Because the handwritten digit dataset is unbalanced, the data needs to be balanced using Synthetic Minority Over-Sampling Technique (SMOTE). The CLEO algorithm finds the best parameter values for SVM to operate in classification. It indicates the central role of the CLEO algorithm in optimizing the SVM model so that it tends to have the best evaluation results. Assigning an appropriate model in solving a case is an absolute necessity so that it can identify the data pattern and predictions on new data can be made. The hybrid model between the CLEO algorithm and the SVM classifier in this paper has obtained accuracy results for recognizing handwritten digits of 93%. This model was then used to compare between digits. However, the results were entirely satisfactory.
- 14 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 185-189
Analysis of Finding Optimal Current in Electric Circuit Using Chimpanzee Leader Election Optimization
The performance of a metaheuristic algorithm is a challenge in its implementation to obtain convergence in a case. The Chimpanzee Leader Election Optimization (CLEO) algorithm is a metaheuristic model that is relatively new and only relies on the ability of random values to get the best solution. This paper explores CLEO's ability to obtain the best results in the case of a simple electrical circuit consisting of a power supply, resistor, and diode. This technique of obtaining the objective function of a simple electrical circuit deserves attention. Even though it is not something new, at least the reader will get a view that is sufficient to understand the context of how to obtain a minimum function that is equal to zero. The CLEO algorithm, which only relies on random values from the program, is not effective enough to find the optimum convergent value, so the learning rate value is introduced in this model even though it does not change the working principle of this algorithm. The results are significantly different when searching for current from a simple electrical circuit until the error is around 10 −9 from the actual value.
- 14 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 185-189
Analysis of Finding Optimal Current in Electric Circuit Using Chimpanzee Leader Election Optimization
The performance of a metaheuristic algorithm is a challenge in its implementation to obtain convergence in a case. The Chimpanzee Leader Election Optimization (CLEO) algorithm is a metaheuristic model that is relatively new and only relies on the ability of random values to get the best solution. This paper explores CLEO's ability to obtain the best results in the case of a simple electrical circuit consisting of a power supply, resistor, and diode. This technique of obtaining the objective function of a simple electrical circuit deserves attention. Even though it is not something new, at least the reader will get a view that is sufficient to understand the context of how to obtain a minimum function that is equal to zero. The CLEO algorithm, which only relies on random values from the program, is not effective enough to find the optimum convergent value, so the learning rate value is introduced in this model even though it does not change the working principle of this algorithm. The results are significantly different when searching for current from a simple electrical circuit until the error is around 10 −9 from the actual value.
- 8 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 491-496
Classification And Clustering Performances on The Protocol of The Wireless Sensor Network - Chimpanzee Leader Election Optimization
Energy consumption in wireless sensor networks (WSN) is an important thing to note. The nature of real-time data transmission on the WSN forces the WSN to have a shorter lifespan. The existibility of WSN certainly requires a strategy that supports it. This paper aims to observe the performance of a wireless sensor network protocol - chimpanzee leader selection optimization (WSN-CLEO). This protocol has been shown to perform better than some hierarchical routing models at sink positions in the WSN field. The clustering of sensor nodes in this WSN consists of two models, namely clustering and classification. Clustering sensor nodes use an algorithm of chimpanzee leader election optimization (CLEO), while classification uses a threshold value of the distance between the sensor nodes and the base station (BS). However, in the grouping of classification, the threshold distance between the sensor nodes and the BS needs to be proven more efficient in WSN energy management. This comparison method uses several distance classification settings on the WSN -CLEO at 60 meters, 80 meters, and 100 meters. After grouping the sensor nodes, the next step is selecting the group leader. The results of the study show insignificant differences in energy efficiency between the first node dead (FND), half node dead (HND), and last node dead (LND) in the comparison of each scenario. However, they have different tendencies, especially in FND.
- 8 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 491-496
Classification And Clustering Performances on The Protocol of The Wireless Sensor Network - Chimpanzee Leader Election Optimization
Energy consumption in wireless sensor networks (WSN) is an important thing to note. The nature of real-time data transmission on the WSN forces the WSN to have a shorter lifespan. The existibility of WSN certainly requires a strategy that supports it. This paper aims to observe the performance of a wireless sensor network protocol - chimpanzee leader selection optimization (WSN-CLEO). This protocol has been shown to perform better than some hierarchical routing models at sink positions in the WSN field. The clustering of sensor nodes in this WSN consists of two models, namely clustering and classification. Clustering sensor nodes use an algorithm of chimpanzee leader election optimization (CLEO), while classification uses a threshold value of the distance between the sensor nodes and the base station (BS). However, in the grouping of classification, the threshold distance between the sensor nodes and the BS needs to be proven more efficient in WSN energy management. This comparison method uses several distance classification settings on the WSN -CLEO at 60 meters, 80 meters, and 100 meters. After grouping the sensor nodes, the next step is selecting the group leader. The results of the study show insignificant differences in energy efficiency between the first node dead (FND), half node dead (HND), and last node dead (LND) in the comparison of each scenario. However, they have different tendencies, especially in FND.
- 16 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 178-183
Detection of Fake News and Hoaxes on Information from Web Scraping using Classifier Methods
Current technological developments can make humans get information from the hand through gadget media. However, bad and good impacts are indeed a problem that arises in implementing this technology media. Fake news and hoaxes have developed along with social media applications obtained from these technological media. This paper aims to detect fake news and hoaxes using classification modeling. The classification models implemented in this paper are support vector machine (SVM), random forest, nearest centroid, stochastic gradient descent (SGD) method, decision tree (Tree), bagging, AdaBoost, gradient boosting, multi-layer perceptron artificial neural network (MLP ANN), and K-nearest neighbors (K-NN). The data obtained through web scraping amounted to 1116 data from Indonesian language news, with the distribution of training data and test data for modeling of 70% and 30%. The testing data are 335 data consisting of 205 fake news and hoax data and 130 real news data. Web data content processing using the principle of natural language processing (NLP) methods. The random forest model is the best model for classifying fake news and hoaxes with an accuracy value of 89%. The following models with the next high scores are SVM, Gradient Boosting, AdaBoost, SGD, and Decision Tree, respectively, with the highest scores above 80%. In comparison, other methods have accuracy scores below 80%.
- 16 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 178-183
Detection of Fake News and Hoaxes on Information from Web Scraping using Classifier Methods
Current technological developments can make humans get information from the hand through gadget media. However, bad and good impacts are indeed a problem that arises in implementing this technology media. Fake news and hoaxes have developed along with social media applications obtained from these technological media. This paper aims to detect fake news and hoaxes using classification modeling. The classification models implemented in this paper are support vector machine (SVM), random forest, nearest centroid, stochastic gradient descent (SGD) method, decision tree (Tree), bagging, AdaBoost, gradient boosting, multi-layer perceptron artificial neural network (MLP ANN), and K-nearest neighbors (K-NN). The data obtained through web scraping amounted to 1116 data from Indonesian language news, with the distribution of training data and test data for modeling of 70% and 30%. The testing data are 335 data consisting of 205 fake news and hoax data and 130 real news data. Web data content processing using the principle of natural language processing (NLP) methods. The random forest model is the best model for classifying fake news and hoaxes with an accuracy value of 89%. The following models with the next high scores are SVM, Gradient Boosting, AdaBoost, SGD, and Decision Tree, respectively, with the highest scores above 80%. In comparison, other methods have accuracy scores below 80%.
- 16 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 173-177
Classification of Lung Opacity, COVID-19, and Pneumonia from Chest Radiography Images Based on Convolutional Neural Networks
Detection of normal and abnormal lung images from the human chest through media scans in the form of computed tomography (CT) scans or radiographs has received attention for early diagnosis of patients. However, even in the study, the diagnosis obtained better accuracy from CT scan results to detect lungs infected with coronavirus disease 2019 (COVID-19) than a swab test with the polymerase chain reaction (PCR) method. This paper aims to detect and classify normal lungs, lung opacities, lungs infected with COVID-19, and viral pneumonia (in this paper is more commonly written as pneumonia) from human chest radiography images. This paper uses 5000 image data consisting of 2461 typical lung images, 1347 lung opacity images, 295 pneumonia images, and 897 COVID-19 images. The method used to detect and classify the labeled images uses the convolutional neural networks (CNN) method. Several image detection and classification studies often implement this method. Comparing the image data for training and testing data uses a ratio of 80 and 20, respectively. Accuracy results for data training got 99.825%, while data testing got 82.6%.
- 16 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 173-177
Classification of Lung Opacity, COVID-19, and Pneumonia from Chest Radiography Images Based on Convolutional Neural Networks
Detection of normal and abnormal lung images from the human chest through media scans in the form of computed tomography (CT) scans or radiographs has received attention for early diagnosis of patients. However, even in the study, the diagnosis obtained better accuracy from CT scan results to detect lungs infected with coronavirus disease 2019 (COVID-19) than a swab test with the polymerase chain reaction (PCR) method. This paper aims to detect and classify normal lungs, lung opacities, lungs infected with COVID-19, and viral pneumonia (in this paper is more commonly written as pneumonia) from human chest radiography images. This paper uses 5000 image data consisting of 2461 typical lung images, 1347 lung opacity images, 295 pneumonia images, and 897 COVID-19 images. The method used to detect and classify the labeled images uses the convolutional neural networks (CNN) method. Several image detection and classification studies often implement this method. Comparing the image data for training and testing data uses a ratio of 80 and 20, respectively. Accuracy results for data training got 99.825%, while data testing got 82.6%.