Telematika
Not a member yet
319 research outputs found
Sort by
Quality Of Service (QoS) Analysis to Calculate Internet Network Performance Level DISKOMINFOTIK and OPD P3AP2KB Office Riau Province
Full text linkPurpose: Among government agencies, the Office of Communication, Information and Statistics (DISKOMINFOTIK) which functions as an internet Support for Regional Apparatus Organizations (OPD), one of which is the OPD Office for Women\u27s Empowerment, Child Protection, Population Control and Family Planning (P3AP2KB), this study was to determine the quality of the DISKOMINFOTIK network for OPD. P3AP2KB Office.Design/methodology/approach: Analyzing the quality of the internet network at DISKOMINFOTIK and in the P3AP2KB Office OPD room which has a lot of clients during busy hours, free hours, and quiet hours using the Quality Of Service (QoS) method with Throughput, packet loss, delay, TIPHON standard jitter and quantitative methodology.Findings/result: The results of the research prove that the parameter values of Throughput, packet loss, delay are at index 4 with the TIPHON measurement standard in the "Very Good" category, and jitter is index 3 with the TIPHON measurement standard in the "Good" QoS category.Originality/value/state of the art: There has been no previous research to calculate the quality of the internet network provided by DISKOMINFOTIK to Regional Apparatus Organizations (OPD), one of which is the OPD Dinas P3AP2KB, therefore this research was conducted.
Analysis of Factors Affecting Intention to Use and User Satisfaction of Paylater Using DeLone & McLean Adoption Model
Full text linkPurpose: This study aims to determine the factors that affect the intention to use and satisfaction of GoPayLater users in Yogyakarta, by assessing the relationship between variables so that recommendation for improvent can be given.Design/methodology/approach: This study uses the DeLone & McLean adoption model by Seddon which includes 5 constructs namely system quality, information quality, perceived usefulness, intention to use and user satisfaction. Primary data collection was conducted by distributing questionnaires using likert scale measurement to 128 GoPayLater users. The data analysis technique used is SEM-PLS to test the measurement model, structural model and test the hypothesis via SmartPLS software.Findings/results:Based on the results of hypothesis testing in this study, two hypotheses were rejected from eight hypothesises. These findings indicate that perceived usefulness has a positive and significant effect on intention to use, while the variables of system quality and information quality do not have a significant effect directly on intention to use GoPayLater. The R-Square test results show that system quality, information quality and perceived usefulness simultaneously have an effect of 34,4% on intention to use GoPayLater. This study also proves that variables of system quality, information quality and perceived usefulness have a positive and significant effect on GoPayLater user satisfacion, with the level of influence given simultaneously is 51,7% .Originality/value/state of the art: Several previous studies have tested GoPayLater from various aspects, but no research has been found that assesses the relationship and effect of system quality, information quality and perceived usefulness on intention to use and user satisfaction using the DeLone & McLean adoption model by Seddon.
Monitoring Development Board based on InfluxDB and Grafana
Full text linkPurpose: Designing a sensor data monitoring system using a time series database and monitoring platform on a Development Board device.Design/methodology/approach: It begins with a requirement analysis, such as the preparation of the required software and hardware, followed by the creation of the system architecture that will be adopted. Then the development process from a predetermined design to the testing process to ensure the dashboard page can display data according to a predetermined scenario.Findings/result: From the research that has been done, produces a design of sensor data that is sent using the MQTT protocol via Node-RED, then stored in a time series database (InfluxDB) and displayed on the Grafana dashboard display.Originality/value/state of the art: Sensor data monitoring dashboard on Development Board device
Performance Evaluation of Online Smart Parking System in Jakarta
Full text linkPurpose: This research aims to evaluate the performance of the Online Smart Parking System application for 7 consecutive days from 06.00 to 18.00 to evaluate the features in the application, evaluate the level of compliance of the clerks with the use of the application and the level of constraints achievement or fulfillment of each of the rights and obligations of the parties, and identification of the level of achievement of revenue targets.Design/methodology/approach: This research was carried out through several stages which include Product Quality Evaluation, Usage Quality Evaluation, Collaboration Evaluation and Financial Evaluation or financial aspects.Findings/result: Design of Logical Framework Application System Online Smart Parking System.Originality/value/state of the art: This research focuses on evaluating the design results of the Online Smart Parking System application which is managed by UP Parking Department of Transportation DKI Jakarta with 3 partner applicators
Digital Image Processing to Detect Cracks in Buildings Using Naïve Bayes Algorithm (Case Study: Faculty of Engineering, Halu Oleo University)
Full text linkPurpose: To detect cracks in the walls of buildings using digital image processing and the Naïve Bayes Algorithm.Design/methodology/approach: Using the YCbCr color model for the segmentation process and the HSV color model for the feature extraction process. This study also uses the Naïve Bayes Algorithm to calculate the probability of feature similarity between testing data and training data.Findings/result: Detecting cracks is an important task to check the condition of the structure. Manual testing is a recognized method of crack detection. In manual testing, crack sketches are prepared by hand and deviation states are recorded. Because the manual approach relies heavily on the knowledge and experience of experts, it lacks objectivity in quantitative analysis. In addition, the manual method takes quite a lot of time. Instead of the manual method, this research proposes digital-based crack detection by utilizing image processing. This study uses an intelligent model based on image processing techniques that have been processed in the HSV color space. In addition, this study also uses the YcbCr color space for feature extraction and classification using the Naïve Bayes Algorithm for crack detection analysis on building walls. The accuracy of the research test data reached 88.888888888888890%, while the training data achieved an accuracy of 93.333333333333330%.Originality/value/state of the art: This study has the same focus as previous research, namely detecting cracks in building walls, but has different methods and is implemented in case studies
Preprocessing Using SMOTE and K-Means for Classification by Logistic Regression on Pima Indian Diabetes Dataset
No full textPurpose: Our study aims to combine pre-processing methods to develop a training data model from the Indian diabetic Pima dataset so that it can improve the performance of machine learning in recognizing diabetesDesign/methodology/approach: This research was started through several stages such as collecting the Pima indian diabetes dataset, pre-processing including k-means clustering, oversampling using SMOTE, then undersampling the dataset whose cluster is a minority in each class. Furthermore, the dataset is classified using machine learning namely logistic regression through 10 cross validationFindings/result: The results of this classification performance show that the accuracy reaches 99.5% and is higher than the method in previous studies.Originality/value/state of the art:The method in this study uses SMOTE to handle data imbalances and k-means clustering to remove outliers by removing labels that do not match the majority cluster in each class so that clean data is produced and validation using logistic regression is more accurate than previous studies.Tujuan: Penelitian ini bertujuan untuk menerapkan metode pre-processing untuk membentuk model data latih dari dataset Pima Indian diabetes sehingga dapat meningkatkan performa mesin pembelajaran dalam mengenali diabetes.Perancangan/metode/pendekatan: Riset ini dimulai melalui beberapa tahap yakni pengumpulan dataset Pima Indian diabetes, pre-processing meliputi clustering, oversampling menggunakan SMOTE, kemudian undersampling pada dataset pada klaster minoritas pada setiap kelas. Selanjutnya dataset diklasifikasikan menggunakan machine learning yakni metode regresi logistik melalui 10 cross validationHasil: Hasil dari performa klasifikasi ini menunjukkan akurasi mencapai 99,5% dan lebih tinggi daripada metode pada penelitian sebelumnya.Keaslian/ state of the art: Metode dalam penelitian ini menggunakan SMOTE untuk menangani ketidakseimbangan data dan k-means klastering untuk membuang outlier dengan cara menghapus label yang tidak sesuai dengan klaster mayoritas pada setiap kelas sehingga dihasilkan data yang bersih dan pada validasi menggunakan logistic regression lebih akurat daripada penelitian sebelumnya
Sentiment Analysis Of Student Opinion Related To Online Learning Using Naïve Bayes Classifier Algorithm And SVM With Adaboost On Twitter Social Media
Full text linkTwitter is one of the social media that functions to express opinions on issues or problems that are currently happening, such as problems in the social, economic, educational and other fields. One of the issues being discussed so far is online learning. The government has issued a policy, one of which is for all students to study at home online by using a network to be able to interact with each other like in the classroom. The government\u27s reason for issuing this policy is to break the chain of the spread of the Covid-19 virus, which until now has not subsided. Regarding this online learning policy, there are pros and cons. This opinion is widely expressed on social media, one of which is Twitter. Sentiment analysis is a method for analyzing an opinion which aims to classify texts. The Naïve Bayes Classifier and Support Vector Machine methods are methods machine learning that can be used for sentiment analysis. The problem in classifying text is that the resulting accuracy is less than optimal, so feature selection or boosting is needed to improve its accuracy. In this study, optimization of boosting was carried out using Adaboost. The purpose of this study is to compare the performance of the algorithm before and after using Adaboost. The results of the sentiment analysis on online learning obtained the highest accuracy results by the Naïve Bayes Classifier algorithm coupled with Adaboost of 99.26%, with a precision of 99.39% and recall of 99.20%
Input Variable Selection for Oil Palm Plantation Productivity Prediction Model
Full text linkPurpose: This study aims to implement and improve a wrapper-type Input Variable Selection (IVS) to the prediction model of oil palm production utilizing oil palm expert knowledge criteria and distance-based data sensitivity criteria in order to measure cost-saving in laboratory leaf and soil sample testing.Methodology: The proposed approach consists of IVS process, searching the best prediction model based on the selected variables, and analyzing the cost-saving in laboratory leaf and soil sample testing.Findings/result: The proposed method managed to effectively choose 7 from 19 variables and achieve 81.47% saving from total laboratory sample testing cost.Value: This result has the potential to help small stakeholder oil palm planter to reduce the cost of laboratory testing without losing important information from their plantation
Design Automatic Parking Application of Amikom Purwokerto University
Full text link Purpose: This study aims to deal with parking problems in the area of Amikom University, Purwokerto. In addition, this research is designed to implement theoretical and practical knowledge that has been obtained in lectures.Design/methodology/approach: In research on parking design applications in the Amikom University area, Purwokerto, library study methods and literature study methods are used. The amount of data can add insight and can make it easier to process data in research.Findings/result: This application will be able to help more Amikom Purwokerto University residents, especially in the Faculty of Computer Science. The use of this application will help find parking areas in FIK areas such as Basement Parking, Front Parking and Field Parking. In addition, security will be helped by this application because if it is implemented, vehicles parked in the reserved area will be tidier and safer. In addition, security does not need to find an empty parking area for users.Originality/value/state of the art: This research focuses on parking system design like previous studies. However, this research focuses more on designing parking applications at Amikom Purwokerto University.
Implementation of Mel-Frequency Cepstral Coefficient as Feature Extraction using K-Nearest Neighbor for Emotion Detection Based on Voice Intonation
Full text linkPurpose: To determine emotions based on voice intonation by implementing MFCC as a feature extraction method and KNN as an emotion detection method.Design/methodology/approach: In this study, the data used was downloaded from several video podcasts on YouTube. Some of the methods used in this study are pitch shifting for data augmentation, MFCC for feature extraction on audio data, basic statistics for taking the mean, median, min, max, standard deviation for each coefficient, Min max scaler for the normalization process and KNN for the method classification.Findings/result: Because testing is carried out separately for each gender, there are two classification models. In the male model, the highest accuracy was obtained at 88.8% and is included in the good fit model. In the female model, the highest accuracy was obtained at 92.5%, but the model was unable to correctly classify emotions in the new data. This condition is called overfitting. After testing, the cause of this condition was because the pitch shifting augmentation process of one tone in women was unable to solve the problem of the training data size being too small and not containing enough data samples to accurately represent all possible input data values.Originality/value/state of the art: The research data used in this study has never been used in previous studies because the research data is obtained by downloading from Youtube and then processed until the data is ready to be used for research