Articles
Prediksi Kepuasan Pelanggan dengan Algoritma Rough Set
Breinda, Engla;
Defit, Sarjon;
Nurcahyo, Gunadi Widi
Jurasik (Jurnal Riset Sistem Informasi dan Teknik Informatika) Vol 9, No 1 (2024): Edisi Februari
Publisher : STIKOM Tunas Bangsa Pematangsiantar
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.30645/jurasik.v9i1.735
Bukittinggi, located in West Sumatra Province, hosts approximately 25 computer shops scattered across its various areas. Statistics reveal a proportional distribution of one computer shop per square kilometer within the city limits, intensifying the competition among these establishments. The primary objective of this study is to assess customer satisfaction using the Rough Set Method. Maintaining high levels of customer satisfaction is crucial as it often leads to repeat purchases. The Rough Set Method, renowned for its effectiveness in Knowledge Discovery in Databases (KDD), comprises five key stages: Decision System, Equivalence Class, Discernibility Matrix, Discernibility Matrix Modulo D, Reduction, and General Rule. The dataset utilized in this research originates from HBC Computer Shop in Bukittinggi, comprising records of 96 customers. Through the analysis, a total of 257 rules were generated, facilitating the identification of customer satisfaction levels. Consequently, the findings of this study can serve as valuable insights for HBC Computer Store management in devising marketing strategies to uphold customer satisfaction and effectively compete with similar businesses.
Backpropagation Neural Network Untuk Prediksi Kebutuhan Pemakaian Obat (Kasus Di RSUD dr. Adnaan WD)
Hazlita, H;
Defit, Sarjon;
Nurcahyo, Gunadi Widi
Jurasik (Jurnal Riset Sistem Informasi dan Teknik Informatika) Vol 9, No 1 (2024): Edisi Februari
Publisher : STIKOM Tunas Bangsa Pematangsiantar
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.30645/jurasik.v9i1.736
Artificial Intelligence which is developing increasingly rapidly makes it possible to make predictions. Predictions are made using one of the Artificial Intelligence systems, namely Artificial Neural Networks. Predicting the need for drug use is a problem currently being faced by RSUD dr. Adnaan WD Payakumbuh so that the service is not optimal. This research aims to design an Artificial Neural Network architecture and determine the resulting level of accuracy in predicting the need for drug use. The method used in this research is the Backpropagation method. The stages in the Backpropagation algorithm include the initial weight initialization process, activation stage, weight change and iteration stage. The data processed in this research is drug use data obtained from the Pharmacy Installation at dr. Adnaan WD Payakumbuh Hospital. The results of this research show that the best network architecture is 12-12-1 with a relatively small Mean Squared Error (MSE) value of 0.00685, a Mean Absolute Percentage Error (MAPE) value of 0.1696% and a high level of accuracy reaching 99 .83% for the prediction of Paracetamol 150 mg. The results of this research can help health service centers optimize their services
Penerapan Metode Rough Set Dalam Memprediksi Penjualan Pada PT. Jaya Framex Bengkulu
Lubis, Fitri Amelia Sari;
Lubis, Siti Sahara;
Agustin, Riris;
Karmanita, Deti;
Defit, Sarjon
Jurasik (Jurnal Riset Sistem Informasi dan Teknik Informatika) Vol 9, No 1 (2024): Edisi Februari
Publisher : STIKOM Tunas Bangsa Pematangsiantar
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.30645/jurasik.v9i1.758
So far, in predicting sales at PT. Jaya Framex Bengkulu, only relies on manual calculations. There are no calculations that use a system to help predict sales at PT. Jaya Framex Bengkulu in the future. As more and more entrepreneurs emerge, it requires entrepreneurs to plan sales strategies. So that what is produced does not decrease further, and is not less competitive with other entrepreneurs, to avoid this, it is necessary to have sales predictions to predict sales so that you can plan future sales strategies. Based on the research conducted, the author can draw the conclusion that predicting the number of food products using Data Mining is very helpful in processing data that has been classified such as product supply, product type and capabilities so that it produces rules that support a decision which can later be used as support for sales prediction decisions. to be more optimal. From 13 sample data of the Data Mining sales process using the rough set method, 5 Reducts were produced which were extracted into knowledge of 11 Generate Rules, thereby producing a decision that was conveyed from the resulting rules. The results of this research can be used by developers to predict future sales. It is hoped that adding new variables can produce more varied decisions and more useful knowledge as decision support
Penerapan Metode TOPSIS Untuk Pemberian Bantuan Bedah Rumah Di Nagari Lunang Selatan
Fitriyani, Intan Nur;
Defit, Sarjon;
Nurcahyo, Gunadi Widi
Jurasik (Jurnal Riset Sistem Informasi dan Teknik Informatika) Vol 9, No 1 (2024): Edisi Februari
Publisher : STIKOM Tunas Bangsa Pematangsiantar
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.30645/jurasik.v9i1.738
Indonesian government seeks to improve people's welfare by holding various poverty reduction programs, one of which is providing assistance to uninhabitable houses (RTLH). Equitable development of the welfare of Indonesian society must be comprehensive and even, starting from the smallest scope, namely the village. One of the villages in Indonesia that has implemented a program to provide assistance for uninhabitable houses is Nagari Lunang Selatan which is located in Lunang sub-district, Pesisir Selatan Regency, West Sumatra Province. The implementation of the uninhabitable housing assistance program in Nagari Lunang Selatan has so far still used a manual system so it is not effective because the final results are not objective. There are 5 criteria and 10 alternatives as sample data used in this research. These criteria include the number of dependents, total expenses, total income, land ownership status, and condition of the house. For this reason, this research provides a solution by implementing a decision support system for providing assistance for uninhabitable housing using the Technique For Order of Preference by Similarity to Ideal Solution method, known as TOPSIS, the TOPSIS method is suitable for solving semi-structural problems such as the problem of providing assistance for inadequate housing. inhabit. The aim of this research is to produce a system that can facilitate decision making regarding providing assistance for uninhabitable housing. The results obtained from the test calculation process on sample data of 10 alternatives with 5 criteria provide accurate results. From this test, the results obtained for 3 alternatives as recipients of house renovation assistance
Rancang Bangun Sistem Pakar Backward Chaining Untuk Antisipasi Hama Tumbuhan Kedelai
Resnawita, R;
Siregar, Diffri Solihin;
Adawiyah, Quratih;
Defit, Sarjon
Jurasik (Jurnal Riset Sistem Informasi dan Teknik Informatika) Vol 9, No 1 (2024): Edisi Februari
Publisher : STIKOM Tunas Bangsa Pematangsiantar
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.30645/jurasik.v9i1.759
Soybeans are an important source of protein in Indonesia. Soybean cultivation in Indonesia faces difficulties posed by climate-related elements that create an ideal environment for the proliferation of various pest species, including those that attack leaves and pods. An expert system is a computing system designed to replicate all aspects of expert capacity in decision making. The Backward Chaining method is an approach where the reasoning process begins with the goal or conclusion to be achieved. This research aims to build a system to provide anticipation in dealing with soybean plant diseases and pests. The results of this expert system research support consultations by uploading photos as a consultation medium. Administrators have the ability to manage soybean crop data in the knowledge base, including adding, editing, and deleting data. The expert system is able to display diagnoses accompanied by complete prevention and management solutions for identified symptoms
Metode Multi Attribute Utility Theory (MAUT) Untuk Penilaian Kinerja Guru
Yamin, Abdul Yamin;
Defit, Sarjon;
Sumijan, Sumijan
Computer Science and Information Technology Vol 4 No 3 (2023): Jurnal Computer Science and Information Technology (CoSciTech)
Publisher : Universitas Muhammadiyah Riau
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.37859/coscitech.v4i3.5920
The performance assessment of teachers is a foundation or basis for the development decisions in terms of promotion and career of teachers in a madrasah or school. Currently, teacher performance assessment at Pondok Pesantren MTI Canduang is limited to teachers who are civil servants (PNS) or have obtained certification. In an effort to improve the quality of education, it is important to evaluate the performance of all teachers, including those who are not civil servants. The conventional method of assessment using paper-based evaluation sheets is considered inaccurate and inefficient due to the large number of teachers being assessed. Furthermore, there is no appropriate method for making decisions regarding teacher reward programs. Therefore, the purpose of this research is to apply the Multi Attribute Utility Theory (MAUT) method for teacher performance assessment. This method aims to provide a basis for decision-making in recommending teachers who deserve rewards in each assessment period. Based on the test results using the MAUT method with 40 teacher data and 12 defined assessment criteria, it was found that 3 data points for Tsanawiyah level had the highest value of 0.797 and the lowest value of 0.332, while 3 data points for Aliyah level had the highest value of 0.874 and the lowest value of 0.386. Thus, the research results can help the madrasah determine the best alternatives according to predefined criteria and weights. The resulting web-based application can facilitate the assessment process by making it easier, faster, and more accurate.
Implementasi Naïve Bayes dalam M-Series 4 Mobile Legends untuk Prediksi Kemenangan
Tamaza, Muhammad Abyanda;
Defit, Sarjon;
Sumijan, Sumijan
Computer Science and Information Technology Vol 5 No 1 (2024): Jurnal Computer Science and Information Technology (CoSciTech)
Publisher : Universitas Muhammadiyah Riau
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.37859/coscitech.v5i1.6707
Mobile Legends is a game made by a developer from China called Moontoon which implements the Multiplayer Online Battle Arena (MOBA) system which is currently popular. The popularity of this game is proven by the holding of low, middle and high level tournaments. Recently a high level or international tournament called the M-Series World Championship was held in Indonesia. This game is played by two teams consisting of five players with the aim of destroying enemy targets in the form of towers. The problem in this game is winning and losing. One of the factors that determines victory or defeat is the choice of hero. The wrong hero composition during the draft pick stage can make it difficult for your team to play and lead to unexpected results. This research aims to predict the percentage level of Mobile Legends wins based on the drafted heroes. Prediction is the process of minimizing errors in systematically estimating the future based on past information. The technique used in this research is the Naïve Bayes algorithm. The Naïve Bayes algorithm is a classification method based on probability. This method consists of four stages, namely data understanding, data preparation, data analysis, and results analysis. This research dataset is provided by Youtube MPL Indonesia. The dataset consists of 880 training data and 90 test data for M-Series 4 Mobile Legends. The results of this research provide a percentage value in the form of prediction of 96.67%, precision of 95.65% and recall of 97.78%. The results of an accuracy rate of 96.67% using the Naïve Bayes algorithm show that predictions using the Naïve Bayes algorithm can be applied to predict win ratios in M-Series 4 Mobile Legends.
Penerapan Metode Fuzzy Logic Dalam Sistem Pemantauan Tanaman Berbasis Internet Of Things (Iot) Dengan Arduino
sabil, Muhammad;
Sarjon Defit;
Gunadi Widi Nurcahyo
Computer Science and Information Technology Vol 5 No 1 (2024): Jurnal Computer Science and Information Technology (CoSciTech)
Publisher : Universitas Muhammadiyah Riau
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.37859/coscitech.v5i1.6710
Hydroponic plants in this increasingly modern era, people are increasingly aware that their vegetable needs must be met so that the body's nutritional balance can be met properly. One of the Urban Farming that is suitable in urban areas with narrow dominant land is the hydroponic system. Hydroponics comes from two Greek syllables combined, namely hydro which means water and ponos which means work, so hydroponics means working using air. One of the advantages of this agricultural system is the minimal use of land, where even small areas of land can be utilized. well. Hydroponics is agricultural cultivation without using soil, so hydroponics is an agricultural activity that is carried out using air as a medium to replace soil. Hydroponic systems are increasingly popular among farmers and agricultural service providers because they are able to produce healthier and more productive plants without using soil as a growing medium. This research aims to test the performance of an Internet of Things (IoT) based Hydroponic Monitoring System using Arduino on plants or vegetables with the method used in this research is Fuzzy logic. This method has 3 stages, namely Fuzzification, Defuzzification, Fuzzy Rule. The data set processed in this research was taken from measurements of pH and temperature on hydroponic vegetable plants in the PKK garden of Kemantan Kebalai Village. The dataset consists of 340 data. The results of this research can identify and calculate the percentage of pH and temperature measurements with an accuracy level of 90%. Therefore, this research can be a reference in measuring acid, normal and alkaline levels in hydroponic plants.
Implementasi Data Mining untuk Pemetaan Persebaran Infeksi Human Imunodeficiency Virus di Provinsi Riau
Fadillah, Riszki;
Sarjon Defit;
Sumijan
Computer Science and Information Technology Vol 5 No 1 (2024): Jurnal Computer Science and Information Technology (CoSciTech)
Publisher : Universitas Muhammadiyah Riau
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.37859/coscitech.v5i1.6712
Based on data released by the Riau Provincial Health Service until October 2022, there were 8034 people living with HIV/AIDS (PLWHA), of which 3,711 were in the AIDS stage. Human Immunodeficiency Virus is a virus that attacks the body's immune system, while Acquired ImmunoDeficiency Syndrome (AIDS) is a collection of diseases caused by the HIV virus due to damage to the immune system in humans, resulting in the body being susceptible to potential diseases. This research aims to map the spread of HIV/AIDS in Riau Province to prevent and control the spread of the HIV/AIDS virus by the relevant agencies. The method used in this research is Fuzzy C-Means to carry out clustering in districts/cities which will then be visualized using a map or with a Geography Informatics System (GIS). The Fuzzy C-Means method is a data grouping technique that uses the existence of each data point in A cluster as determined by the degree of membership. The output from Fuzzy C-Means is a series of cluster centers and several degrees of membership for each data point. The data used in this research is HIV/AIDS data in Riau Province from 1997 to 2023. Based on the results of the tests that have been carried out, the results obtained are 3 clusters, namely the safe zone has 5 districts/cities, the alert zone has 5 districts/cities, and There are 2 districts/cities in the dangerous zone. There needs to be treatment through the Health Service, the AIDS Control Commission, and related Non-Governmental Organizations (NGOs) to prevent and control HIV/AIDS in Riau Province for areas that have a high potential for the spread of HIV/AIDS. The tests that have been carried out obtain a minimum error value of 0.008251 in the 8th iteration with the performance of Fuzzy C-Means being 13.271 in the distance between clusters.
Penerapan Convolutional Neural Network pada Klasifikasi Citra Pola Kain Tenun Melayu
Mukhlis Santoso;
Sarjon Defit;
Yuhandri
Computer Science and Information Technology Vol 5 No 1 (2024): Jurnal Computer Science and Information Technology (CoSciTech)
Publisher : Universitas Muhammadiyah Riau
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.37859/coscitech.v5i1.6713
The use of electronic computerized media is growing along with advances in hardware and software as an analytical tool with various algorithms and methods for classifying and measuring objects in various contexts. This progress aims to overcome the weaknesses that exist in conventional methods used in the identification process. The identification process can be applied to various objects, one of which is an image object. An image is a visual representation of an object formed through a combination of RGB (red, green, blue) colors. RGB color components or features have a range of values from 0 to 255 in an image. Weaving is a type of fabric that is specially made with distinctive motifs. Malay weaving motifs have a lot of diversity, this diversity makes it difficult to distinguish the motifs of these fabrics.This study aims to recognize and distinguish the pattern of Malay woven fabric. The method used in this research is Convolutional Neural Network (CNN). The CNN method has several stages, namely Convolution Layer, Pooling Layer, Rectifed Linear Unit (ReLU) Function, Fully-Connected Layer, Transfer Learning, Optimizer and Accuracy. The dataset used in this research is sourced from Tenun Putri Mas Bengkalis. The dataset used consists of 1000 images of weaving motifs which are divided into 80% training data and 20% testing data, from the existing dataset divided into three categories of weaving motifs namely pucuk rebung, elbow clouds and elbow keluang. The results in this study are considered good because they produce accuracy with a result of 95% with an epoch value of 15. From the results of good enough accuracy, it is hoped that it can help the community in recognizing Malay weaving motifs.