Jurnal Sistem Informasi dan Ilmu Komputer

Application Of Yolo V8 For Product Defect Detection In Manufacturing Companies

2024-08-03T03:47:03+00:00

One important aspect in the production process is maintaining product quality and avoiding defects that could harm the company. This research aims to improve quality and avoid product defects that are detrimental to the company, especially defects in the form of bubbles in the product, by using YOLOv8. The dataset consists of 100 data which is divided into 80 for training and 20 testing data with an epoch value of 100. To obtain optimal bubble detection results, this research chose the latest version of YOLOv8 and compared several models, namely YOLOv8n, YOLOv8s, YOLOv8m, YOLOv8l, and YOLOv8x. The research results show that YOLOv8m achieves the highest accuracy among other models with a mAP value of 0.712, precision of 0.764, recall of 0.659, and F1-score of 0.708. This research highlights the potential of detection models that can detect bubbles precisely and accurately.

Keywords: Kecacatan Produk, Deteksi Gelembung, Perusahaan Manufaktur, Model YOLOv8

Sentiment Analysis of Public Opinions Regarding "Ideas of Presidential Candidates" in YouTube Video Comments with Robustly Optimized BERT Pretraining Approach

2024-07-09T02:59:03+00:00

Social media and video-sharing platforms such as YouTube have become one of the primary sources of information and social interaction in modern society. In politics, YouTube has become essential for spreading ideas, campaign platforms, and opinions about the presidential election. Using the pre-trained Indonesian Roberta Base Sentiment Classifier Model, the data obtained from YouTube comments will be divided into three labels: positive, negative, and neutral. The results of this study are the accuracy for each sentiment label, where the value for positive is 93%, the negative is 90.5%, and the neutral is 93.04%. Residents give more positive comments to presidential candidate Prabowo Subianto, with a positive value of 54.13%, followed by Anies Baswedan at 42.8% and Ganjar Pranowo at 31.91%.

IMPLEMENTATION OF SUPPORT VECTOR MACHINE AND HARMONY SEARCH FOR CATARACT SEVERITY CLASSIFICATION IN FUNDUS IMAGES

2024-08-19T02:54:38+00:00

Cataract is a condition that causes clouding of the lens of the eye and is a leading cause of blindness, including in Indonesia. Cataract diagnosis is often inconsistent between ophthalmologists due to personal experience. This research proposes a Support Vector Machine (SVM) based classification system and Harmony Search metaheuristic algorithm to optimize the weight vector 'w' on the SVM hyperplane as a supporting tool for cataract diagnosis. The research data comes from Kaggle which includes normal eye fundus images and cataracts with mild-moderate and severe levels. The research stages include image conversion from RGB to Grayscale, image enhancement with Histogram Equalization and GLCE, and feature extraction using GLCM and Haar Wavelet Transform, and unbalanced data is balanced by the SMOTEENN method. The results showed that Harmony Search successfully improved SVM accuracy compared to Conventional SVM using Gradient Descent. Accuracy increased by 18% from 0.53 to 0.71 on unbalanced data, and by 13% from 0.67 to 0.80 on balanced data. In addition, Harmony Search can improve computational time efficiency due to its ability to explore space globally.

OPTIMIZATION OF NODE SIZE CONFIGURATION IN CNN-ELM MODEL FOR BRAIN TUMOR MRI IMAGE CLASSIFICATION

2024-08-21T06:57:19+00:00

This study proposed a method to classify four types of brain tumors—Glioma, Meningioma, Pituitary, and Non-Tumor—using the Kaggle Brain Tumor MRI Dataset. The research involved stages of data collection, preprocessing, model design, model training, and evaluation. A hybrid Convolutional Neural Network - Extreme Learning Machine (CNN-ELM) algorithm was employed, demonstrating the importance of selecting the optimal number of hidden nodes for achieving high accuracy. The test results revealed that with 2000 hidden nodes, the CNN-ELM model achieved an overall accuracy of 98.86%, with F1-scores of 97% for Glioma, 98% for Meningioma, 100% for Non-Tumor, and 100% for Pituitary tumors. In comparison, the model with 1000 hidden nodes achieved an accuracy of 96.96%, while models with 3000 and 4000 hidden nodes achieved 98.10% and 96.58% accuracy, respectively. These findings highlight the critical role of hidden node selection in optimizing model performance. The CNN-ELM algorithm proves to be a viable alternative for classifying brain tumor MRI images, contributing to advancements in medical technology.

Herbal Leaf Image Classification Using Convolutional Neural Network (CNN)

2024-08-20T10:15:36+00:00

This research delves into the application of Convolutional Neural Networks (CNNs) to address the complexities of identifying herbal leaf species in Indonesia, often challenging due to the vast variations in shape, color, and texture. Utilizing a dataset of herbal leaf images acquired using the Bing Downloader Scrapping technique, a CNN model was trained to classify various plant varieties with a remarkable accuracy rate of 92.66%. Additionally, the analysis of low loss values indicates that the model not only effectively maps the intricate features of each image to the correct category but also efficiently reduces error rates. These findings offer a significant contribution to the context of herbal medicine development and biodiversity conservation, opening up avenues for technological integration in efforts to preserve Indonesia's natural and cultural resources.

Grouping Diseases of Patients at RSU Mitra Medika Bandar Khalippa Medan Using the K-Medoids Clustering Method

2024-08-21T02:23:18+00:00

The aim of this research is to apply the methodK-Medoisin categorizing the illnesses of patients at the RSUBandar Khalippa Medika PartnersMedan. And to produce a system for grouping patient data based on Rapidminer and Google Colab on patient diseases.Based on the results of research on the application of the K-Medoids algorithm, it was found that the grouping of patient diseases at RSU Mitra Medika Bandar Khalippa used the RapidMiner application with a C0 (High) cluster of 3 diseases, a C1 (Medium) cluster of 6 diseases and a C2 (Low) cluster of 1 disease. Meanwhile, using the Google Colabs application with a C0 (High) cluster of 3 diseases, a C1 (Medium) cluster of 4 diseases and a C2 (Low) cluster of 3 diseases. The results of grouping patient disease data at RSU Mitra Medika Bandar Khalippa using RapidMiner, it was found that the disease with the highest grouping (C0)is a diseasePulmonary tuberculosis, Essential Hypertension and Diabetes Mellitus. Whereasgrouping patient disease datawith Google Colabs it was found that the disease with the highest grouping (C0)is a diseaseBronchus Or Lung, Trachea, Bronchus And Lung and Pleural Effusion.

Keywords: Disease Grouping, RSU, MethodsK-Medois.

Utilization Of Discord Bots In Providing Manhwa Recommendations Using Content-Based Filtering Method

2024-08-24T02:04:16+00:00

The high number of manhwa released today makes it difficult for readers to find manhwa that match their preferences, especially when trying to find manhwa that are similar to ones they have read before. The manual search process, either through recommendations from communities or online forums, often results in subjective and inconsistent suggestions. To address this issue, a Discord bot was developed that utilizes the Content-Based Filtering method as an automated solution to manhwa recommendation. This method uses the Cosine Similarity algorithm to measure the similarity between manhwa based on features such as title, genre, synopsis, and author. For comparison, the Euclidean Distance algorithm is used to evaluate the accuracy and performance of the recommendation. From the test results, the Cosine Similarity algorithm showed superior performance in providing recommendations based on the questionnaire results and showed a high level of user satisfaction with the developed Discord bot.

Analysis of User Behavior of the Digital Korlantas Polri Application with Integrated UTAUT in the Community of East Java Province

2024-08-03T03:34:50+00:00

Currently, more and more agencies are developing their services by utilizing technology in the form of mobile applications, including the Republic of Indonesia Police agency, especially the Traffic Corps with its mobile application-based service called Digital Korlantas Polri. This research aims to identify what factors influence users' interests, behavior and intentions towards the National Police Traffic Corps Digital application by testing 10 variable hypotheses built from the integration of the Unified Theory of Acceptance and Use of Technology (UTAUT), Technology Acceptance Model (TAM), Theory models. of Planned Behavior (TPB) and Service Quality with the educational level factor proposed as a moderator. The results of this research show that Performance Expectancy and Effort Expectancy are the 2 main factors influencing Behavioral Intention. Other results show that factors such as Social Influence, Facilitating Conditions, and Perceived Risk have a negative influence where they can reduce the user's Behavioral Intention. Apart from that, the research results also show that Behavioral Intention and Word of Mouth can influence users to continue using this application and educational level factors have a negative influence on users' behavioral intention to continue using the application on an ongoing basis.

Keywords: Unified Theory of Acceptance and Use of Technology, Theory of Planned Behavior, Technology Acceptance Model, Service

Application of the Rapid Application Development Method to Analyze the MBKM Information System at Prima Indonesia University

2024-08-22T03:48:45+00:00

Independent Campus (MBKM) is an innovation carried out by the Ministry of Education, Culture and Research to produce graduates who are ready to face rapid technological advances and one of the campuses that has implemented it is Universitas Prima Indonesia (UNPRI), where students have difficulty with UNPRI MBKM activities. to take part in MBKM activities due to misinformation and lack of information. To improve the MBKM activity process, researchers will re-analyze the MBKM management information system, namely based on the RAD (Rapid Application Development) model. By implementing the Rapid Application Development (RAD) model in the MBKM program at Prima Indonesia University, it was able to analyze according to the administrator's needs, where the RAD model found that the analysis of business processes from student track records was incomplete, and for further research it was necessary to develop stakeholders who were willing to join with the campus.

Application of Smote and Decision Tree Classification in Detecting Fraudulent Transactions

2024-08-03T03:42:14+00:00

Fraud detection in online transactions is critical to protecting consumers and maintaining the integrity of the online business ecosystem. Dataset imbalance can affect the classification prediction performance. To overcome data imbalance, this research uses an oversampling approach with the SMOTE method. The aim of this research is to analyze the performance of the SMOTE algorithm and decision tree classification in dealing with data imbalance problems in fraudulent transactions. The dataset used is online payments taken from Kaggle. The dataset shows that there are unbalanced classes, and it was found that using the SMOTE method increased the performance value better than using it without the SMOTE method. Using SMOTE gets very high metric values, up to a recall value of 100%. This shows that the model used in classifying fraudulent transactions is very effective.

Visual Attention Analysis of Perspective Images Using the Eye Tracking Method

2024-08-03T03:47:39+00:00

In this research, eye tracking was applied to observe students' visual attention to three perspective images, each of which has a Region of Interest (RoI). In initial research, it was found that the majority of students faced difficulties in concentrating during the learning process. The aim of this research is to analyze the visual attention of perspective drawings in adolescents in an effort to increase learning concentration. Eye tracking was used as a research instrument to monitor the eye movements of 70 students objectively and in real-time who were guided by giving assignments to look for certain objects. This study showed that in terms of perception speed and focus duration, female participants outperformed male participants. However, overall the level of concentration of teenagers cannot be said to be good. These findings provide important knowledge for educators in creating more effective visual content to improve student concentration and understanding.

Classification Of Egg Quality Using The K-Nearest Neighbor Algorithm In Machine Learning

2024-07-30T04:16:21+00:00

In addition to meat, fish, and milk, one of the staple foods consumed by the community is chicken eggs. Egg quality assessment is separated into two categories: exterior (egg shell) and interior (egg contents). However, the evaluation method used in this investigation is focused on evaluating the external quality of eggs. Pre-processing, feature extraction, classification, and evaluation are steps taken in the image processing method used to classify chicken eggs. Classification methods that can be used include the K-Means Clustering and K-Nearest Neighbor (KNN) methods and improved KNN. Based on the findings in the study, the KNN improvisation method can be used to classify chicken egg quality, with a test accuracy value of 91.67%.

Design Of A Detection Device For Dangerous Gas In Electric Cigarette Users Using Arduino Nano

2024-07-19T19:38:19+00:00

The use of electronic cigarettes (vapes) is expected to be a healthier alternative to tobacco cigarettes. However, various studies have shown that electronic cigarettes also contain hazardous substances such as Benzene and Carbon Monoxide which damage lung health. Previous studies have studied methods for detecting hazardous gases in the lungs, but most of them use expensive devices and complex methods such as SPME–GC/MS which are not easily accessible to the public. This study aims to develop a portable, easy-to-use, and affordable, non-invasive and real-time hazardous gas detection device, especially for Carbon Monoxide and Benzene. This device uses an Arduino microcontroller board and MQ-7 and MQ-135 gas sensors to measure gas levels from users' breath. The results showed that 30 participants exceeded the safe limit of Benzene of 0.5 ppm and 17 participants exceeded the safe limit of Carbon Monoxide of 6 ppm, indicating the potential risk to lung health of electronic cigarette users. These findings can increase public awareness of the dangers of electronic cigarettes and encourage reduction or cessation of their use.

Keywords: Gas level detection, Lungs, Electronic cigarettes, Arduino.

A Data Pre-processing Strategy Utilizing Adaptive Masking for the Classification of Pediatric Pneumonia Using VGG-16

2024-08-24T02:11:41+00:00

Pneumonia is still a leading cause of death in children, especially in areas with limited medical resources. This study aims to test several pre-processes to find the best set of pre-processes that can be applied to the children's chest X-ray dataset by applying adaptive masking, histogram equalization, CLAHE and Gaussian blur. Then, childhood pneumonia is classified using a CNN architecture, namely VGG-16. By applying these pre-processing methods, this study is divided into several scenarios. The highest accuracy was obtained from scenario 1, which used a combination of adaptive masking, histogram equalization and Gaussian blur, resulting in an accuracy of 94%. Scenario 2 uses histogram equalization and Gaussian blur with an accuracy of 92%. Then Scenario 3 uses a histogram equalization replacement for CLAHE with a combination of adaptive masking, CLAHE and Gaussian blur with 93% accuracy. Finally, scenario 4 uses a combination of CLAHE and Gaussian blur methods with 91% accuracy. In addition, this research also addresses the challenges posed by unbalanced data sets and the need for highly accurate detection tools.

Implementation of comparison of K-means algorithm with C4.5 algorithm to predict the feasibility of being a Catholic

2024-08-14T05:02:30+00:00

This research conducted a comparison between two methods, namely C4.5 and K-Means, with the aim of improving the work efficiency of the Catholic Church Secretariat in selecting data of prospective Catholic congregants. These methods were developed to assist secretariat employees in determining the eligibility of valid files and data of prospective Catholic congregants through a web-based application. The data used comes from the selection results of several files and data of prospective congregants that have been collected by the Catholic Church Secretariat. Data analysis was carried out using the K-Means and C4.5 algorithms to predict the feasibility of the prospective congregants' files. It is hoped that the results of this research can help the Catholic Church Secretariat to improve work efficiency, both in terms of time and effort, in the selection process of prospective Catholic congregants and increase accuracy in determining the eligibility of the data of each applicant at the Catholic Church Secretariat improve the efficiency of the selection process and enhance the accuracy in determining the suitability of potential members.

Analysis Of Customer Satisfaction With Solaria Restaurants In Medan City Using K-Means Clustering Method

2024-08-29T16:46:18+00:00

Customer satisfaction is the key to the success of a company in the modern business context. In strategic planning and business management, a focus on customer satisfaction has become essential to ensure that the services a company provides not only retain customers but also enable sustainable growth. In the case of Solaria Restaurant, it helps to find groups of customers with comparable satisfaction patterns, which allows businesses to optimize their marketing strategies and improve their service quality. Specifically, this study uses the K-Means Clustering method to evaluate customer satisfaction with Solaria Restaurant services. The research utilized an online questionnaire distributed to 250 surveyed people to assess factors such as service response and the physical appearance of the restaurant that affect customers' perceptions of the business. The results showed that customer satisfaction is generally considered very good, the results of clustering analysis of customer satisfaction at Solaria Restaurant resulted in the number of very good clusters is cluster I. This result increases our understanding of customer preferences. These results increase our understanding of customer preferences and build a basis for improvement strategies that focus more on improving the customer experience.

Analysis Of E-Commerce Systems To Improve Sales Strategy Using Descriptive Methods In The Sarana Jaya Electronic Company

2024-08-03T03:38:58+00:00

E-commerce includes various activities such as sales, procurement, distribution and promotional transactions carried out via the Internet online network or electronic platforms. PT. Sarana Jaya Elektronik is a company that specializes in the distribution of electronic devices. Currently, PT. Sarana Jaya Elektronik has used an e-commerce system, namely the Tokopedia application, to sell its products. In order to find out the effect of implementing an e-commerce system on improving sales strategies, an analysis process can be carried out using descriptive methods. Descriptive research refers to a methodology in which researchers examine events and phenomena in the lives of individuals, encouraging one or a group of individuals to narrate their experiences.

APPLICATION OF NAIVE BAYES ALGORITHM FOR SALES ANALYSIS AT ERIGO STORE

2024-07-14T11:11:24+00:00

The purpose of this study is to research and compare the accuracy of the previous research algorithm, namely the KNN algorithm with the Naive Bayes algorithm, for the evaluation of Erigo Store sales. Given the increasingly fierce market competition, it is very necessary to formulate a marketing strategy to analyze and predict products using data mining processing methods. Data mining is the introduction of patterns, machine learning techniques, statistics, and visualization techniques that aim to provide information to make better decisions and improve prediction accuracy through the process of analyzing data based on the Knowledge Discovery in Database (KDD) procedure. The research dataset was taken from shopee Toko Erigo e-commerce sales data using web scraping techniques, starting from January 2021 to June 2023 consisting of 5 categories of Erigo Store products, namely Shirts, T-Shirt, Outwear, Jacket and Pants. The overall accuracy of the previous research product using the KNN algorithm was 83.62% while the study using the application of the Naive Bayes algorithm for sales analysis in Erigo stores achieved an accuracy of 98.3% by using Matlab to analyze the data. The accuracy of the T-shirt category reached 98.6%, the shirt category reached 98.4%, the pants category reached 98.1%, the outwear category reached 98.7% and the accuracy of the jacket category reached 97.6%.

Utilization Of Website-Based Technology For Analysis And Prevention Of Stunting Using The Fuzzy Tsukamoto Methods

2024-08-24T03:38:56+00:00

Stunting is a chronic malnutrition condition that has significant impact on inhibiting the growth of a child both physically and intellectually. The study aims toto analyze stunts using a web-based technology system using Tsukamoto’s Fuzzy method. This method was chosen for its ability to deal with the uncertainty and variability medical data. The system integrates various variables that affect stunts, such as nutritional intake and physical growth, to produce a more accurate diagnosis. The research was carried out by collecting data from various health sources and applying the fuzzy Tsukamoto method to process the data. The trial subjects in this developmental study were 30 children aged 1–60 months, or 0–5 years. Subjects were selected by random sampling, consisting of 6 children from 1–5 years of age each. Based on the results of the analysis, it appears that the fuzzy Tsukamoto-based system development trial can provide a better prediction of the risk of stunting in children compared to conventional methods. Using this approach, it is expected to help health workers take more accurate steps in the treatment and prevention of stunts.

Keywords: Stunting, Fuzzy Tsukamoto Method, Nutritional Analyisis, Technology Systems, Child Heal

THE IMPACT OF AN ANIMATED VIDEO INQUIRY TRAINING MODEL ON JUNIOR HIGH SCHOOL SCIENCE STUDENTS’ LEARNING OUTCOMES

2024-08-29T03:20:22+00:00

The purpose of this study was to determine how the use of the questioning learning model with animated videos at Dr. Wahidin Sudirohusodo Private Junior High School in Medan, North Sumatra, impacts student learning outcomes on science materials. This study used question training, with two groups used for pretest and posttest. The study involved all students of Dr. Wahidin Sudirohusodo Private Junior High School in Medan, North Sumatra. The study involved grade VIII students spread across several parallel classes. The sample was randomly selected. There were 66 students in class VIII-1 and VIII-3. Class VIII-1 was the experimental class with 34 students, and class VIII-3 was the control class with 32 students. The results of the research show that the Inquiry Training Learning Model using animated video media on the science learning outcomes of students in the experimental class shows a good attitude compared to the control class, as seen from the increase in scores as evidenced by the experimental class’s mean score which is higher compared to the control class with the mean score for the post-test for the experimental class was 65.7647 and for the control class 42.6250.

Implementation Of The ARIMA Method In Predicting LQ 45 Stock Prices (UNTR Issuer)

2024-09-03T07:15:09+00:00

The implementation of technology is used in running businesses or activities that generate profits, such as predicting investments on the stock exchange through transaction data in the transaction data base. Machine learning is an algorithm that produces an approximation function that connects input variables so that it has the potential to be implemented in stock predictions. Stock investment has the characteristics of high risk - high return. Losses are caused by investors' lack of knowledge. Stock value analysis is divided into two, namely fundamental analysis and technical analysis. Technical analysis uses data or records about the market to try to access the demand and supply of a particular stock or the market as a whole. Based on the problems found by investors or bankers, this research will use the autoregressive integrated moving average (ARIMA) method to predict stock price movements. The Arima method consists of four stages, namely identifying time series methods, estimating parameters for alternative methods, testing methods and estimating time series values. Based on these problems, the ARIMA method will be used to predict stock movements. The Arima model (1,0,2) with RMS: 2200.576849857124 successfully predicted for the next 180 days

Analysis And Prediction Of Global Population Using Random Forest Regression

2024-08-13T01:49:32+00:00

This research evaluates the performance of the random forest regression algorithm in predicting global population growth from time series data. The findings indicate that population growth predictions remain stable, with an annual increase of less than 1%. Model analysis using evaluation metrics such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and model scores demonstrates high quality, with average values below 0.5. These results imply that the model can deliver optimal and consistent outcomes. The model shows potential for accurate predictions when tested on datasets. Further analysis reveals a population increase of 0.88% in 2024, equating to an addition of approximately 70,206,291 people, and a rise of 0.91% in 2025, adding about 73,524,552 people

Implementation of UI/UX Using Design Thinking Method in Tomsufood Application

2024-09-04T07:46:04+00:00

Tomsufood is one of the online food ordering applications. To provide the best experience in ordering food, Tomsufood designed a new application product. An approach called Design Thinking was used in this study. The Design Thinking method is a software product design approach that is based on innovation and relies on problem-solving techniques. After identifying and understanding the problem obtained through the process of identifying the problem, describing the solution, empathizing with the user, and prototyping and testing. So that the Tomsufood application is able to solve problems that occur in society. The results of the study showed that the application of the Design Thinking method was able to produce a UI/UX design that was more intuitive and responsive to user needs. Users reported increased satisfaction in using the application, which included ease of navigation, clarity of information, and efficiency in the food ordering process.

Keywords: UI/UX, Design Thinking, Tomsufood, User Experience

Analysis and Design of Web-Based MSME Cashier Application Using the Waterfall Model

2024-08-03T03:43:05+00:00

The rapid advancement of information technology in Indonesia has provided numerous benefits such as increased work efficiency, expanded access to information, and enhanced connectivity. Additionally, technology has opened up new opportunities for business and job creation. However, on the other hand, there are several challenges that need to be addressed, such as digital divide, cybercrime, and social impacts that require management. Manual transaction systems have several drawbacks, particularly during high order volumes, resulting in inefficient recording and difficulty in tracking top-selling products. This is due to the limitations of manual sales transaction recording. Therefore, the use of a cashier application is necessary to streamline all transaction processes, including sales, payments, purchases, payroll, and inventory management. Point of Sale (POS) is a software and hardware solution specifically designed to facilitate these transactions. Despite the availability of many POS applications in the current market, they often do not meet the specific needs of SMEs and are relatively expensive, leading many SME operators to hesitate in adopting them. Based on observations and interviews with SME operators, there is a need for POS applications that not only cater to the specific requirements of SMEs but are also financially feasible.

Application Of Support Vector Machine Method To Predict Heart Disease

2024-09-01T14:34:14+00:00

Heart attack disease is when the arteries are blocked by fatty deposits This results in symptoms like chest discomfort and dyspnea. Furthermore, Damage to the heart muscle can result from obstructed or reduced blood flow to the heart. Heart attack disease remains Indonesia’s greatest cause of death as of right now. The current problem is that it is very difficult to predict heart disease and identify heart disease. The right method is needed to predict heart disease. The purpose of this study was to calculate the level of accuracy of the Support Vector Machine method in predicting heart attack disease. The research findings and data analysis conducted utilizing the Support Vector Machine algorithm yielded an accuracy rate of 91.8%. Thus, it can be said that in comparison to the K-Nearest Neighbor approach, the support vector machine algorithm is superior in predicting the development of heart attack disease, which achieved an accuracy of 88%, and Logistic Regression, which achieved 83% accuracy.

Keywords: Heart Attack, Support Vector Machine, Prediction.

Analysis of Student Satisfaction with Services Using the Service Quality (ServQual) Method

2024-08-28T04:41:06+00:00

This study aims to analyze the level of student satisfaction with services at SMK Swasta Perguruan Mabar, a private vocational school in North Sumatra that offers education in Accounting. The quality of school services, which includes learning services, management, facilities, and other aspects, is the main focus in an effort to increase student satisfaction and improve the quality of education. The Service Quality (ServQual) method is used in this study to evaluate service quality based on five dimensions: Tangibles, Reliability, Responsiveness, Assurance, and Empathy. The results of the analysis show that the level of student satisfaction is 3.482, which is categorized as good. This finding indicates that the ServQual method is effective in measuring student satisfaction by using expected value as the main benchmark, thus providing more accurate and objective results in assessing student satisfaction with the services provided. This research is expected to provide insight for school management to further improve the quality of services provided to students.

Keywords: Students, Satisfaction, School, Service Quality

Analysis And Design Of A Survey Application To Measure Outpatient Loyalty Using The Customer Relationship Management Method

2024-08-24T03:25:14+00:00

Addressing patient complaints regarding the services provided indicates potential issues with service quality, possibly even non-compliance with procedures. To identify such issues, regular surveys must be conducted. Conducting surveys requires specialized human resources and significant costs, which can be burdensome for hospitals. Therefore, a software application was developed to conduct surveys and process the data, allowing results to be analyzed and used for policy-making. This software uses the Customer Relationship Management (CRM) method. The software was tested for 6 months at RSU Royal Prima. In the first month, many service quality issues were identified, and SOPs were not followed correctly. However, in the following months, these issues were addressed, and after 6 months of testing, the survey software proved to be beneficial with relatively low costs, as it did not require specialized human resources. In other words, the CRM survey software can enhance transparency, responsiveness, and service quality, as well as manage patient satisfaction levels without relying on third parties.

Comparison of K-Nearest Neighbors and Convolutional Neural Network Algorithms in Potato Leaf Disease Classification

2024-08-03T03:45:04+00:00

tatang.rohana@ubpkarawang.ac.id3, santi.arum@ubpkarawang.ac.id4, deden.wahiddin@ubpkarawang.ac.id5
ABSTRACT
Potato production in Central Java was recorded to have decreased by 10.77% by the Central Statistics Agency (BPS), from 278,717 tons in 2022 to 248,700 tons in 2023. This decline is due to the fact that potatoes are susceptible to diseases such as late blight and dry spot (early blight) which can significantly reduce yields. This study aims to evaluate the performance of Convolutional Neural Network (CNN) with VGG16 architecture and K-Nearest Neighbors (KNN) to find the best method for potato late blight classification. The dataset used consists of 1500 potato leaf images divided into training, validation, and testing. This research uses pre- processing including resizing, rescaling, and data augmentation. The results show that CNN with the VGG16 model is superior in classifying potato leaf diseases compared to KNN with the MobileNetV2 model. CNN produced an accuracy of 96% while KNN with the MobileNetV2 model obtained an accuracy of 93%. These results can be used as a powerful tool in supporting potato leaf disease identification. This model makes a significant contribution to the development of disease identification techniques through digital image processing.
Keywords: Potato Leaf Disease, Convolutional Neural Network, VGG16, K-Nearest

Analysis of the Management Information System of MBKM at Prima Indonesia University Using the Waterfall Method

2024-07-26T04:03:24+00:00

Implementing MBKM activities in the field has challenges for students and the unavailability of MBKM information within the campus environment is an obstacle to participating in MBKM. This research is to analyze and build a web-based MBKM information system using the Laravel framework and an effective System Development Life Cycle (SDLC) model to support the implementation and monitoring of MBKM activities on campus. Researchers provide solutions to the obstacles faced by students in participating in MBKM activities, so that this program can run more smoothly and achieve the researchers' goals. The SDLC model is a model for developing website-based information systems. SDLC provides a systematic and structured approach to ensure that the software being developed meets requirements. The research results show that the SLDC model produces business processes that need to be developed, namely adding access rights for stakeholders to make it easier to assess study programs, and the need for involvement of system programs (Prodi) for assessing MBKM student activities. Researcher's suggestions for the security of using user accounts and access rights for Human Resources with experience in MBKM activities.

Data Mining Analysis In Minimizing Company Losses Using Fuzzy Time Series Method

2024-08-03T03:35:51+00:00

Losses are the most avoided by all business entities in this case the research obtained a research study at PT. Sumatera Sarana Sekar Sakti. The company suffered a big loss in the expenditure / spending section that was not managed properly. The existence of excess funds or shortages in each company's expenditure is a form of loss, not only in the form of material but even immaterial. Therefore, this research conducts an analysis by generating data predictions so that a value is obtained that will minimize company losses because it provides the right and efficient funds. The method used in prediction is Fuzzy Time Series. It is a new category of methods that have been widely used in various studies because they produce good predictive values. In this study, the Fuzzy Time Series method produces 0.82% error rate from data analysis of 1875 company expenditure transactions. Measurement of the prediction error rate using Mean Absolute Percentage Error which is often called MAPE. It is a measurement that is often used in various studies with data prediction categories.