Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
Not a member yet
1071 research outputs found
Sort by
Comparing the Performance of Data Mining Algorithms in Predicting Sentiments on Twitter
Get PDFOn Twitter, users can post tweets, videos, and images. It can, however, also be disruptive and difficult. To categorize the material and improve searchability, hashtags are crucial. This study focuses on examining the opinions of Twitter users who participate in trending topics. The algorithms K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) are used for sentiment analysis. The data set comprises tweet information on popular topics that was collected using the Twitter API and saved in Excel format. SVM and K-NN are used for data preparation, weighting, and sentiment analysis. With 105 data points, the study provides insight into user sentiment. SVM identified 99% of positive responses and 1% of negative responses with an accuracy of 80%. KNN successfully identified 90% of the positive responses and 10% of the negative responses, with an accuracy rate of 71.4%. According to the results, SVM performs better when analyzing the sentiment of hashtag users on Twitter.
On the social networking site Twitter, users can post tweets, videos, and images. It can, however, also be disruptive and difficult. In order to categorize material and improve searchability, hashtags are crucial. This study focuses on examining the opinions of Twitter users who participate in trending topics. The algorithms K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) are employed for sentiment analysis. The dataset comprises of tweet information on popular subjects that was collected using the Twitter API and saved in Excel format. SVM and K-NN are used for data preparation, weighting, and sentiment analysis. With 105 data points, the study provides insights into user sentiment. SVM identified 99% of positive and 1% of negative replies with accuracy of 80%. KNN successfully identified 90% of positive and 10% of negative responses, with an accuracy rate of 71.4%. According to the results, SVM performs better when analyzing the sentiment of hashtag users on Twitter
Pengembangan Aplikasi Mobile untuk Penyelesaian Vehicle Routing Problem
No full textThe vehicle routing problem (VRP) is a combinatorial optimization problem faced by transportation services related to pick up or delivery, such as industrial raw materials distribution, tour and travel, or travel routing problems in general. VRP is an NP-hard problem where the higher the dimensions of the problem will have a higher computational complexity. Without realizing it, VRP problem are often encountered every day. Therefore, it will be very useful if VRP solver is implemented in mobile application media. So, the aim of this work is developing a mobile application to get the shortest path and minimal cost in VRP problem. It is integrated by both Mapbox API and Google Maps API to get a real distance for modeling problem. The result show that the developed application can run well in all possibility condition
Credit Scoring Model for Farmers using Random Forest
Get PDFOne of the problems faced by farmers in Indonesia is capital. Based on Indonesian Central Statistics Agency survey results, the number of farmers who borrow capital from formal institutions such as banks is still small. This is because the process of applying for loans at banks is lengthy, farmers are considered high-risk and unbankable, and the rating of the agricultural sector is unattractive to banks. This study aims to determine the attributes and design a model of agricultural credit assessment. This study uses secondary data related to bank credit ratings and land productivity from banks in the Telagasari sub-district in 2018–2020 and Cipayung sub-district in 2020. Data were analyzed using random forests. The research process includes four stages: data collection, data pre-processing, model building, and model analysis and evaluation. This study produced five important variables that are relevant to farmers: planting costs, sales, land productivity, total production, and land area. The model built produces the most optimal accuracy of 83% with an AUC score of 81%. Based on the AUC performance classification, it can be concluded that the model that has been made is good at predicting the credit status of farmers because the AUC value is included in the good classification predicate.
One of the problems faced by farmers in Indonesia is capital. Based on Indonesian Central Statistics Agency survey results, the number of farmers who borrow capital from formal institutions such as banks is still small. This is because the process of applying for loans at banks is lengthy, farmers are considered high-risk and unbankable, and the rating of the agricultural sector is unattractive to banks. This study aims to determine the attributes and design a model of agricultural credit assessment. This study uses secondary data related to bank credit ratings and land productivity from banks in the Telagasari sub-district in 2018–2020 and Cipayung sub-district in 2020. Data were analyzed using random forests. The research process includes four stages: data collection, data pre-processing, model building, and model analysis and evaluation. This study produced five important variables that are relevant to farmers: planting costs, sales, land productivity, total production, and land area. The model built produces the most optimal accuracy of 83% with an AUC score of 81%. Based on the AUC performance classification, it can be concluded that the model that has been made is good at predicting the credit status of farmers because the AUC value is included in the good classification predicate
Implementing Agile Scrum Methodology in The Development of SICITRA Mobile Application
Get PDFSoftware development management system is important in application development. A proper software development management system will create a team that can adapt to system requirements and changes during application development. Various software development management systems are developed and widely implemented in software development, one of which is Agile Scrum. This study aims to implement well-documented Scrum for end-to-end application development, including the development of servers and mobile applications that we develop. We developed a bus application called SICITRA, with the main feature of being able to help passengers share their travel information with those closest to them. Scrum is used because it has agility which can make application development faster and more organized, and there is a close relationship between everyone involved in the project. The results of this study are that by using well-documented Scrum, we can make it easier to track progress, become a guide during system development, become history and evaluate Scrum implementation during development.Software development management system is important in application development. A proper software development management system will create a team that can adapt to system requirements and changes during application development. Various software development management systems are developed and widely implemented in software development, one of which is Agile Scrum. This study aims to implement well-documented Scrum for end-to-end application development, including the development of servers and mobile applications that we develop. We developed a bus application called SICITRA, with the main feature of being able to help passengers share their travel information with those closest to them. Scrum is used because it has agility which can make application development faster and more organized, and there is a close relationship between everyone involved in the project. The results of this study are that by using well-documented Scrum, we can make it easier to track progress, become a guide during system development, become history and evaluate Scrum implementation during development
Software as a Service-based Integrated Interactive Online Course System
Get PDFSeveral online learning platforms use recorded videos as a medium for delivering their material. In addition, several applications, such as Zoom, Google Meet, and WhatsApp, can help communicate interactively during this learning process. However, because the application is still separate from the existing online learning platform, users must switch applications and make the necessary data settings. Another obstacle experienced is that not all online learning service providers can have the infrastructure to use online learning systems, especially if the providers are individuals. This research uses a simple sequential method and aims to build an integrated online course system equipped with an interactive learning management system that many online courses, in general, can use. The main features are submissions of new courses by teachers, ordering courses by students, classes can be held using live video streams, the interaction between teachers and students in real-time with live chat, and a learning management system that includes sending and receiving assignments and quizzes. The test results show that as many as 73.2% of respondents gave the highest score for the built system.
Several online learning platforms use recorded videos as a medium for delivering their material. In addition, several applications, such as Zoom, Google Meet, and WhatsApp, can help communicate interactively during this learning process. However, because the application is still separate from the existing online learning platform, users must switch applications and make the necessary data settings. Another obstacle experienced is that not all online learning service providers can have the infrastructure to use online learning systems, especially if the providers are individuals. This research uses a simple sequential method and aims to build an integrated online course system equipped with an interactive learning management system that many online courses, in general, can use. The main features are submissions of new courses by teachers, ordering courses by students, classes can be held using live video streams, the interaction between teachers and students in real-time with live chat, and a learning management system that includes sending and receiving assignments and quizzes. The test results show that as many as 73.2% of respondents gave the highest score for the built system
Image Transformation With Lung Image Thresholding and Segmentation Method
No full textImage transformation is important to obtain and find certain information about an image that was not previously known, such as pixels, geometry, size, and color. Following this, this research aims to analyze image transformation in producing better values using threshold and segmentation methods. The segmentation process is carried out based on two color models, namely hue saturation value (HSV) and red green blue (RGB). The image data used in this study was the x-ray image of the lungs from www.fk.unair.ac.id. which is processed using the Matlab 2021a application to help the analysis process. on the results of the image segmentation analysis carried out in this case, the greater the HSV and RGB threshold values used in the image data, the better and clearer the segmentation of the detected image results. In other words, the size of the thresholding value generated greatly affects the quality, brightness, size, and color of the resulting image. The best lung X-ray image segmentation results were obtained when using the threshold values HSV = 0.9 and RGB = 9.
Image transformation is important to obtain and find certain information about an image that was not previously known, such as pixels, geometry, size, and color. Following this, this research aims to analyze image transformation in producing better values using threshold and segmentation methods. The segmentation process is carried out based on two color models, namely hue saturation value (HSV) and red green blue (RGB). The image data used in this study was the x-ray image of the lungs from www.fk.unair.ac.id. which is processed using the Matlab 2021a application to help the analysis process. on the results of the image segmentation analysis carried out in this case, the greater the HSV and RGB threshold values used in the image data, the better and clearer the segmentation of the detected image results. In other words, the size of the thresholding value generated greatly affects the quality, brightness, size, and color of the resulting image. The best lung X-ray image segmentation results were obtained when using the threshold values HSV = 0.9 and RGB = 9
Image Convolution to Obtain Color ROI after Segmentation Process with Fuzzy Cmeans
Get PDFImage segmentation is still an important concern in terms of digital image processing. Segmentation refers to dividing an image into several parts based on similar characteristics or uniformity. Its use is quite important, especially related to the analysis and application of digital image processing. The challenge faced is separating the object image from its background in images with complex backgrounds. The aim of this research is to separate tomatoes from simple to complex backgrounds. This paper proposes a convolution method of segmented binary images and RBG images all based on contours using Fuzzy C-means and reconstruction operations to obtain the foreground from an image with a complex background. This method has been tested on ripe tomatoes with various backgrounds. This method has Indicated Performance Achievement Sc = 99.2%, Fpe = 0.6% and FNe = 0.4%. This shows that the method is suitable and robust for the dataset used in this study, especially if it will be continued for further work related to the classification of tomato maturity assessment.Image segmentation is still an important concern in terms of digital image processing. Segmentation refers to dividing an image into several parts based on similar characteristics or uniformity. Its use is quite important, especially related to the analysis and application of digital image processing. The challenge faced is separating the object image from its background in images with complex backgrounds. The aim of this research is to separate tomatoes from simple to complex backgrounds. This paper proposes a convolution method of segmented binary images and RBG images all based on contours using Fuzzy C-means and reconstruction operations to obtain the foreground from an image with a complex background. This method has been tested on ripe tomatoes with various backgrounds. This method has Indicated Performance Achievement Sc = 99.2%, Fpe = 0.6% and FNe = 0.4%. This shows that the method is suitable and robust for the dataset used in this study, especially if it will be continued for further work related to the classification of tomato maturity assessment
Comparison of Naive Bayes and PSO-Based Naive Bayes Algorithms for Prediction of Covid-19 Patient Recovery Data in Indonesia
Get PDFA brand new disease known as COVID 19 was identified in 2019 but has yet to infect humans (World Health Organization, 2019). This group of viruses can infect mammals, including humans and birds, and cause sickness. People commonly contract coronaviruses from the flu and other minor respiratory diseases, but they can also spread serious diseases such as SARS, MERS, and the deadly COVID-19. Therefore, to avoid further casualties, this number must be decreased. It is crucial to understand the variables that can truly reduce the danger of death and gauge the propensity for recovery in Covid-19 patients. Several techniques in data mining can be used to forecast patient recovery rates depending on various characteristics. The criteria of this study included gender, age, province, and status. The Naive Bayes (NB) and Pso-based Naive Bayes algorithms are compared in this study using patient data sets to determine whether the strategy is more accurate. The findings of this study reveal that the NB method has a 94.07% accuracy rate, a precision value of 14%, a recall value of 1% and an AUC value of 0.613, according to the study data. The accuracy rate of the Naive Bayes based on PSO is 95.56%, the precision is 25%, the recall is 1%, and the AUC is 0.540.
A brand-new illness known as COVID 19 was identified in 2019 but has yet to infect humans (World Health Organization, 2019). This group of viruses can infect mammals including humans as well as birds and cause sickness. People commonly contract coronaviruses from the flu and other minor respiratory ailments, but they can also spread serious diseases like SARS, MERS, and the deadly COVID-19. So that there are no more casualties, this number must be decreased. It is crucial to understand the variables that can truly reduce the danger of death and gauge the propensity for recovery in Covid-19 patients. Several techniques in data mining can be used to forecast patient recovery rates depending on various characteristics. This study's criteria included gender, age, province, and status. The Naive Bayes (NB) and Pso-based Naive Bayes algorithms are compared in this study using patient datasets to determine whether strategy is more accurate. The findings of this study reveal that using the NB method has a 94.07% accuracy rate, a precision value of 14%, a recall value of 1%, and an AUC value of 0.613, according to the study's data. The accuracy rate of PSO-based Naive Bayes is 95.56%, the precision is 25%, the recall is 1%, and the AUC is 0.540
Performance Comparison of Convolutional Neural Network and MobileNetV2 for Chili Diseases Classification
Get PDFChili is an important agricultural commodity in Indonesia and plays an significant role in the economic growth of the country. Its demand from households and industries reaches up to 61%. However, this high demand also means that monitoring efforts must be intensified, particularly for chili plant diseases that can greatly impact yields. If these diseases are not addressed promptly, they can lead to a decrease in production levels, which can negatively affect the economy. With technological advancements, automatic monitoring using image processing is now highly feasible, making monitoring more efficient and effective. Common chili plant diseases include chili leaf yellowing disease, chili leaf curling disease, cercospora leaf spots, and magnesium deficiency with symptoms that can be observed through the shape and color of the leaves. This research aims to classify chili plant diseases by comparing the CNN algorithm and the pre-trained MobileNetV2 based model performance using the Confussion Matrix. The study shows that the MobileNetV2 model, trained with a learning rate of 0.001, produces a more optimal model with an accuracy of 90% and based on the calculation of the confusion matrix, the average percentage values for recall, precision, and F1 score are 92%. These findings highlight the potential.Chili is an important agricultural commodity in Indonesia and plays a significant role in the nation's economic growth. Its demand by households and industries reaches up to 61%. However, this high demand also means that monitoring efforts need to be intensified, particularly for chili plant diseases that can greatly impact yields. If these diseases are not promptly addressed, they can lead to a decrease in production levels, which can negatively affect the economy. With technological advancements, automatic monitoring using image processing is now highly feasible, making monitoring more efficient and effective. Common chili plant diseases include Chili leaf yellowing disease, Chili leaf curling disease, and cercospora leaf spots and Magnesium Deficiency with symptoms that can be observed through the shape and color of the leaves. This research aims to classify chili plant diseases by comparing the CNN algorithm and the pre-trained MobileNetV2 based model performance using Confussion Matrix. The study shows that the MobileNetV2 model, trained with a learning rate of 0.001, produces a more optimal model with an accuracy of 90% and based on the calculation of the confusion matrix, the average percentage values for recall, precision, and F1 score are 92%. These findings highlight the potentia
Q-Madaline: Madaline Based On Qubit
Get PDFThis research focuses on developing the MADALINE algorithm using quantum computing. Quantum computing uses binary numbers 0 or 1 or a combination of 0 and 1. The main problem in this research is how to find other alternatives to the MADALINE algorithm to solve pattern recognition problems with a quantum computing approach. The data used in this study are heart failure data to predict whether a patient is at risk of death. The data source comes from KAGGLE, consisting of 299 data with 12 symptoms and one target, alive or dead. The result of this study is an alternative to the MADALINE algorithm that uses quantum computing. The precision of the test results with MADALINE with a learning rate of 0.1 = 100% with 2 epochs. The accuracy of the test results using a quantum approach with a learning rate of 0.1 is 85.71%. The results of this study can be an alternative to the MADALINE algorithm with a quantum computing approach, although it has not shown better accuracy than the classical MADALINE algorithm. More research is needed to produce better accuracy with larger data