Jurnal Infotel (Sekolah Tinggi Teknologi Telematika Telkom Purwokerto)
Not a member yet
392 research outputs found
Sort by
Construction of cardiac arrhythmia prediction model using deep learning and gradient boosting
Full text linkArrhythmia is a condition in which the rhythm of heartbeat becomes irregular. This condition in extreme cases can lead to fatal heart attack accidents. In order to reduce heart attack risk, appropriate early treatments should be conducted right after getting results of Arrhythmia condition, which is generated by electrocardiography ECG tools. However, reading ECG results should be done by qualified medical staff in order to diagnose the existence of arrhythmia accurately. This paper proposes a deep learning algorithm method to classify and detect the existence of arrhythmia from ECG reading. Our proposed method relies on Convolutional Neural Network (CNN) to extract feature from a single lead ECG signal and also Gradient Boosting algorithm to predict the final outcome of single lead ECG reading. This method achieved the accuracy of 96.18% and minimized the number of parameters used in CNN Layer.Arrhythmia is a condition in which the rhythm of heartbeat becomes irregular. This condition in extreme cases can lead to fatal heart attack accidents. In order to reduce heart attack risk, appropriate early treatments should be conducted right after getting results of Arrhythmia condition, which is generated by electrocardiography ECG tools. However, reading ECG results should be done by qualified medical staff in order to diagnose the existence of arrhythmia accurately. This paper proposes a deep learning algorithm method to classify and detect the existence of arrhythmia from ECG reading. Our proposed method relies on Convolutional Neural Network (CNN) to extract feature from a single lead ECG signal and also Gradient Boosting algorithm to predict the final outcome of single lead ECG reading. This method achieved the accuracy of 96.18% and minimized the number of parameters used in CNN Layer
A User Recommendation Model for Answering Questions on Brainly Platform
Full text linkBrainly is a Community Question Answer (CQA) application that allows students or parents to ask questions related to their homework. The current mechanism is that users ask questions, then other users who are in the same subject interest can see and answer it. As a reward for answering questions, Brainly gives points. The number of points varies by question. The greater of total points users have, Brainly will automatically display them in the smartest user leaderboard on the site's front page. But sometimes, some users do not have good activity in answering questions. Thus, it is possible to have an urgent question that has not been answered by anyone. This study implements Fuzzy C-Means cluster method to improve Brainly's feature regarding the speed and accuracy of answers. The idea is to create student clusters by utilizing the smartest students' leaderboard, subjects interest, and answering activities. The stages applied in this research started with Data Extraction, Preprocessing, Cluster Process, and User Recommender. The optimal number of clusters in the answerer recommendation in the Brainly platform is 2 clusters. The value of the fuzzy partition coefficient for two clusters reached 0.97 for Mathematics and 0.93 for Indonesian. Meanwhile, the results of the recommendations were influenced by answer ratings. Many numbers of the answer are not given rating because the possibility of the answers are not appropriate or user's insensitivity in giving ratings.Brainly is a Community Question Answer (CQA) application that allows students or parents to ask questions related to their homework. The current mechanism is that users ask questions, then other users who are in the same subject interest can see and answer it. As a reward for answering questions, Brainly gives points. The number of points varies by question. The greater of total points users have, Brainly will automatically display them in the smartest user leaderboard on the site's front page. But sometimes, some users do not have good activity in answering questions. Thus, it is possible to have an urgent question that has not been answered by anyone. This study implements Fuzzy C-Means cluster method to improve Brainly's feature regarding the speed and accuracy of answers. The idea is to create student clusters by utilizing the smartest students' leaderboard, subjects interest, and answering activities. The stages applied in this research started with Data Extraction, Preprocessing, Cluster Process, and User Recommender. The optimal number of clusters in the answerer recommendation in the Brainly platform is 2 clusters. The value of the fuzzy partition coefficient for two clusters reached 0.97 for Mathematics and 0.93 for Indonesian. Meanwhile, the results of the recommendations were influenced by answer ratings. Many numbers of the answer are not given rating because the possibility of the answers are not appropriate or user's insensitivity in giving ratings
Automatic detection of covid-19 based on CT Scan images using the convolution neural network
Full text linkThe 2019 coronavirus pandemic (Covid-19) has been declared a health emergency by WHO with the death rate steadily increasing worldwide, various efforts have been made to deal with this pandemic, from prediction to receiving medical imaging. CT Scan and chest X-Ray images have been proven to be accurate to help medical personnel diagnose COVID, in this paper, we propose a convolutional neural network (CNN) approach and the DenseNet transfer learning model series which aims to understand and find the best classification for COVID or Non-COVID detection. On CT scan chest images, we made two special models in the Descent series, then compared the CNNs in both models by calculating the Accuracy, Precision, Recall, and F1-Score values and presented the results in the confusion matrix. The testing framework is carried out on CNN and the first model of the DenseNet series uses adam optimization, the input function is 244x244x3, the soft-max function is applied as an activity with losses across entropy categories, epoch 50, and batch size for training and testing 16 while validation uses batch size 8, the EarlyStopping function also determined, From the test results, the CNN model is superior to the Densenet series of the first model with an accuracy of about 0.76 (76%), when testing the second model, we carried out the shifting, zooming process and changed the input function to 64x64x3, epoch 30 by adding 4 layers. The second model approach produces better accuracy than CNN and the first DenseNet series, but not as good as expected, based on the test results on the second model produces an accuracy of 0.90 (90%) on Densenet169, Densenet121 around 0.88 (88%) and last Densenet201 is about 0.83 83%), so it is superior to simple CNN model
Design and Implementation of Robotank for Room Monitoring and Exploration
Full text linkA robot is a mechanical device that can perform physical tasks, either autonomously or with human control. Robots began to be used for monitoring in areas that have narrow spaces and/or dangerous areas. So that this robot must be able to carry out monitoring with a remote control system. Therefore, in this study, a robotank is designed that can perform space exploration with remote control. Robotank is designed to use a track and wheel that can pass through various terrains and it has dimensions of 11.8 x 10.8 x 9.1 cm. Robotank is equipped with a camera to monitor in real-time. Robotank can move from one point to another by controlling using a remote control system with a maximum distance of 20 meters in line of sight terrain and 16 meters in non-line of site fields, with an average speed of 0.84 m/s. Robotank can work for 1 hour 52 minutes. With this robotank, it is hoped that it can be used for exploration of areas or rooms that have small spaces and dangerous.A robot is a mechanical device that can perform physical tasks, either autonomously or with human control. Robots began to be used for monitoring in areas that have narrow spaces and/or dangerous areas. So that this robot must be able to carry out monitoring with a remote control system. Therefore, in this study, a robotank is designed that can perform space exploration with remote control. Robotank is designed to use a track and wheel that can pass through various terrains and it has dimensions of 11.8 x 10.8 x 9.1 cm. Robotank is equipped with a camera to monitor in real-time. Robotank can move from one point to another by controlling using a remote control system with a maximum distance of 20 meters in line of sight terrain and 16 meters in non-line of site fields, with an average speed of 0.84 m/s. Robotank can work for 1 hour 52 minutes. With this robotank, it is hoped that it can be used for exploration of areas or rooms that have small spaces and dangerous
Multi-aspect sentiment analysis on netflix application using latent dirichlet allocation and support vector machine methods
Full text linkAmong many film streaming platforms that have sprung up, Netflix is the platform that has the most subscribers compared to the other platforms. However, not all reviews provided by the Netflix users are good reviews. These reviews will later be analyzed to determine what aspects are reviewed by the users based on reviews written on the Google Play Store, using the Latent Dirichlet Allocation (LDA) method. Then, the classification process using the Support Vector Machine (SVM) method will be carried out to determine whether each of these reviews is included in the positive or negative class (Sentiment Analysis). There are 2 scenarios that were carried out in this study. The first scenario resulted that the best number of LDA topics to be used is 40, and the second scenario resulted that the use of filtering process in the preprocessing stage reduces the score of the f1-score. Thus, this study resulted in the best performance score on LDA and SVM testing with 40 topics, and without running the filtering process with the score of 78.15%.Among many film streaming platforms that have sprung up, Netflix is the platform that has the most subscribers compared to the other platforms. However, not all reviews provided by the Netflix users are good reviews. These reviews will later be analyzed to determine what aspects are reviewed by the users based on reviews written on the Google Play Store, using the Latent Dirichlet Allocation (LDA) method. Then, the classification process using the Support Vector Machine (SVM) method will be carried out to determine whether each of these reviews is included in the positive or negative class (Sentiment Analysis). There are 2 scenarios that were carried out in this study. The first scenario resulted that the best number of LDA topics to be used is 40, and the second scenario resulted that the use of filtering process in the preprocessing stage reduces the score of the f1-score. Thus, this study resulted in the best performance score on LDA and SVM testing with 40 topics, and without running the filtering process with the score of 78.15%
The Modelling of Nonlinear Distance Sensor Using Piecewise Newton Polynomial with Vertex Algorithm
Full text linkThe Sharp GP2Y0A02YK0F is categorized as a nonlinear sensor for distance measurement. This sensor is also categorized as a low-cost sensor. The higher resolution, cheap, high accuracy and easy to install are the advantages. The accuracy level of this sensor depends on the type of the measured object materials, requires an additional device unit and further processing is required since the output is non-linear. The distance determination is not easy for this type of sensor since the characteristic of this sensor fulfills non-injective function. The modelling process is one of methods to convert the output voltage of the sensor to a distance unit. The advantages of polynomial modelling are simple form model, moderate in flexibilities of shape, well known and understood properties, and easy to use for computational matters. The obstacle of polynomial-based modelling is the presence of Runge’s phenomenon. The minimization of Runge’s phenomenon can be done with decreasing the model order. The piecewise Newton polynomials with vertex determination method have been succeeded to generate a nonlinear model and minimize the occurrence of Runge’s phenomenon. The low level of MSE by 0.001 and error percentage of 2.38% has been obtained for the generated model. The low MSE level leads to the high accuracy level of the generated model.The Sharp GP2Y0A02YK0F is categorized as a nonlinear sensor for distance measurement. This sensor is also categorized as a low-cost sensor. The higher resolution, cheap, high accuracy and easy to install are the advantages. The accuracy level of this sensor depends on the type of the measured object materials, requires an additional device unit and further processing is required since the output is non-linear. The distance determination is not easy for this type of sensor since the characteristic of this sensor fulfills non-injective function. The modelling process is one of methods to convert the output voltage of the sensor to a distance unit. The advantages of polynomial modelling are simple form model, moderate in flexibilities of shape, well known and understood properties, and easy to use for computational matters. The obstacle of polynomial-based modelling is the presence of Runge’s phenomenon. The minimization of Runge’s phenomenon can be done with decreasing the model order. The piecewise Newton polynomials with vertex determination method have been succeeded to generate a nonlinear model and minimize the occurrence of Runge’s phenomenon. The low level of MSE by 0.001 and error percentage of 2.38% has been obtained for the generated model. The low MSE level leads to the high accuracy level of the generated model
Breast cancer recurrence prediction system using k-nearest neighbor, naïve-bayes, and support vector machine algorithm
Full text linkBreast cancer is a serious disease and one of the most fatal diseases in the world. Statistics show that breast cancer is the second common cancer worldwide with around two million new cases per year. Some research has been done related to breast cancer, and with the advancements of technology, breast cancer can be detected earlier by using artificial intelligence or machine learning. There are popular machine learning algorithms that can be used to predict the existence or recurrence of breast disease, for example, k-Nearest Neighbor (kNN), Naïve Bayes, and Support Vector Machine (SVM). This study aims to check the prediction of breast cancer recurrence using those three algorithms using the dataset available at the University of California, Irvine (UCI). The result shows that the kNN algorithm gives the best result in terms of accuracy to predict breast cancer recurrence.Breast cancer is a serious disease and one of the most fatal diseases in the world. Statistics show that breast cancer is the second common cancer worldwide with around two million new cases per year. Some research has been done related to breast cancer, and with the advancements of technology, breast cancer can be detected earlier by using artificial intelligence or machine learning. There are popular machine learning algorithms that can be used to predict the existence or recurrence of breast disease, for example, k-Nearest Neighbor (kNN), Naïve Bayes, and Support Vector Machine (SVM). This study aims to check the prediction of breast cancer recurrence using those three algorithms using the dataset available at the University of California, Irvine (UCI). The result shows that the kNN algorithm gives the best result in terms of accuracy to predict breast cancer recurrence
Improved RSSI-based path-loss model for indoor positioning and navigation in LabVIEW using trilateration
Full text linkIndoor positioning and navigation now contribute in many applications to track and direct people inside the building. The popular trilateration technique is utilized to detect user’s position through three access point of Bluetooth low energy. However, received signal from Bluetooth has insignificancy due to the noise, multipath, fading or other radio propagation. A study of received signal characteristics in specific indoor locations must be considered to predict and improve the accuracy of estimation. In this case, the adjustment of raw received signal readings is essential. we extracted linear regression model by compare between raw and analytical value of received signal power. Then, utilizing the corrected received signal, finding the best suitable path loss exponent model is required in order to minimize position estimation error. The last step is applying the additional model and the chosen path-loss on LabVIEW as a mean to visualize position and navigation system. The result yield that the new model gives lower error on 2 out of 3 access points. The corresponding path loss exponent n = 2.1 is selected to comply with the indoor environment in this case. The lowest RMSE yields 1.24 and considered as a good level of accuracy. The Navigation system worked well providing route to the desired location in the Laboratory.Indoor positioning and navigation now contribute in many applications to track and direct people inside the building. The popular trilateration technique is utilized to detect user’s position through three access point of Bluetooth low energy. However, received signal from Bluetooth has insignificancy due to the noise, multipath, fading or other radio propagation. A study of received signal characteristics in specific indoor locations must be considered to predict and improve the accuracy of estimation. In this case, the adjustment of raw received signal readings is essential. we extracted linear regression model by compare between raw and analytical value of received signal power. Then, utilizing the corrected received signal, finding the best suitable path loss exponent model is required in order to minimize position estimation error. The last step is applying the additional model and the chosen path-loss on LabVIEW as a mean to visualize position and navigation system. The result yield that the new model gives lower error on 2 out of 3 access points. The corresponding path loss exponent n = 2.1 is selected to comply with the indoor environment in this case. The lowest RMSE yields 1.24 and considered as a good level of accuracy. The Navigation system worked well providing route to the desired location in the Laboratory