AbstrakCase-Based Reasoning (CBR) merupakan sistem penalaran komputer yang menggunakan pengetahuan lama untuk mengatasi masalah baru.CBR memberikan solusi terhadap kasus baru dengan melihat kasus lama yang paling mendekati kasus baru. Hal ini akan sangat bermanfaat karena dapat menghilangkan kebutuhan untuk mengekstrak model seperti yang dibutuhkan oleh sistem berbasis aturan. Penelitian ini mencoba untuk membangun suatu sistem Penalaran Berbasis Kasus untuk melakukan diagnosa penyakit THT (Telinga, Hidung dan Tenggorokan). Proses diagnosa dilakukan dengan cara memasukkan kasus baru (target case) yang berisi gejala-gejala ang akan didiagnosa ke dalam sistem, kemudian sistem akan melakukan proses indexing dengan metode backpropagation untuk memperoleh indeks dari kasus baru tersebut. Setelah memperoleh indeks, sistem selanjutnya melakukan proses perhitungan nilai similarity antara kasus baru dengan basis kasus yang memiliki indeks yang sama menggunakan metode cosine coefficient. Kasus yang diambil adalah kasus dengan nilai similarity paling tinggi. Jika suatu kasus tidak berhasil didiagnosa, maka akan dilakukan revisi kasus oleh pakar. Kasus yang berhasil direvisi akan disimpan ke dalam sistem untuk dijadikan pengetahuan baru bagi sistem. Hasil penelitian menunjukkan sistem penalaran berbasis kasus untuk mendiagnosa penyakit THT ini membantu paramedis dalam melakukan diagnosa. Hasil uji coba sistem terhadap 111 data kasus uji, terdapat 9 kasus yang memiliki nilai similarity di bawah 0.8. Kata kunci—case-based reasoning, indexing, similarity, backpropagation, cosine coefficient Abstract Case-Based Reasoning (CBR) is a reasoning system that uses old knowledge to solve new problem. CBR provides solutions to new cases by looking at old case that comes closest to the new case. It will be very useful because it eliminates the need to extract the model as required by the rule-based systems. This studytriestoestablisha system forCBR for diagnosingdiseasesof ENT.Diagnosisprocessis done byinsertinga new casethat containsthe symptoms ofthe disease to bediagnosed, thenthe system willdo theindexingprocess with backpropagation method toobtainan indexofnewcases. Afterthat, the systemdo thecalculation of the valueof similaritybetweenthe newcasebycasebasiswhichhas thesame indexwithnew cases using cosine coefficient method. The casetaken isthe casewiththe highestsimilarityvalue. If acaseis not successfullydiagnosed, thecasewillbe revisedby theexperts and it can be used asnew knowledgefor thesystem. The results showedcase-basedreasoningsystemtodiagnosediseasesof ENTcan helpparamedicsin performingdiagnostics. The test results of 111 data test cases, obtained 9 cases that have similarity values below 0.8. Keywords—case-based reasoning, indexing, similarity, backpropagation, cosine coefficient
The nutrition problem is very important problem that need more attention. If someone doesn't know about his nutrition status, he can't control how much nutrition value should be needed by his body. In this research, it has been built a Decision Support System (DSS) to compute nutrition status. The system needs physical condition from user via user interface. Pocket PC platform used to develop this DSS. The computation of nutrition status based on K-Nearest Neighbor (K-NN). The K-NN method will look for the shortest distance between evaluated data and K nearest data in the data training setThe result of this research shows that this system can help user to get information about his nutrition status, so he can keep his nutrition normal status to avoid he diseases attack.
The X-ray image is a medical examination procedure that uses electromagnetic wave radiation to get a picture of the inside of the body. However, in the process, there is noise that appears due to the exposure factor. This research builds a system to improve the X-ray image with noise by using Gaussian Filter and Histogram Equalization. In this study, in order to see the optimization of image enhancement, the two methods were combined. The data used are 60 x-ray images that have noise and each has an original image without noise as a comparison image to get system accuracy using PSNR and SSIM. Gaussian Filter method is used to reduce noise by determining the size of the kernel matrix and the standard deviation used. Histogram Equalization method is used to even out the value of the gray level of the image. Based on the test results from the combination of the two methods, the larger the size of the kernel matrix used, the faster the duration of time needed to repair the image. The PSNR value and accuracy obtained in the X-ray image repair are 31 dB and 71% on a 3x3 kernel matrix with an average time duration of 9 seconds, 32 dB and 77% on a 5x5 kernel matrix with an average duration of 9 seconds, 32 dB and 78% on a 7x7 kernel matrix with an average time duration of 8 seconds
Orchid cultivation has been widely carried out by orchid agribusiness in Indonesia, even though there has been a lot of orchid cultivation but there are still obstacles in terms of supplies or maintenance. One of them is water sprinkling and fertilizing the orchid plants which are still done manually. Manual watering is watering done by orchid farmers so it requires a lot of effort and time and the amount of water that is sprinkled is not the same. If the water is poured too much or too little, it can cause rot or dryness of the plant roots so that the plants can die quickly. In the current technological era, watering can be done automatically by utilizing Internet of Things (IoT) technology and implementing the Wireless Sensor Network (WSN) system. NodeMCU ESP32 is used to control all hardware and software components. In this system there are 2 sensor nodes and 1 controller node. Users can control manually or automatically through the website interface. From the results of the implementation and testing it can be concluded that the system is able to provide information on conditions of temperature, air humidity, humidity of the planting medium, water level, liquid fertilizer height and water pH. The system is also capable of running automatic and manual systems, namely controlling and providing on/off condition information on water pumps, fertilizer pumps, faucet 1, faucet 2, faucet 3, fan 1, fan 2 and fan 3 on the website. The effect of the automatic system on the orchid plants was very good, because on the 41st day, the orchid plants using the automatic system experienced faster growth of new shoots, compared to using the manual system. The average delay time for the entire system is 4.5 seconds.
Orange is a type of fruit that is easily found in Sambas Regency. The types that are widely sold are Siam oranges, madu susu and susu. Each type of orange has a different quality and a different price. The price difference often results in fraud committed by traders against buyers to the detriment of the buyer. This is because differentiating types of oranges based on the appearance of the fruit does not have a standard. Therefore, in this study, a citrus fruit classification system was created based on images by implementing deep learning. The method of deep learning used in this research is Convolutional Neural Network (CNN) with AlexNet architecture. The types of oranges that will be observed are madu oranges, madu susu, and siam. The data used are 2250 images of oranges with each class totaling 750 images with a size of 227x227 pixels. The training data is 1575 images and the test data is 675 images. The training is carried out with a total of 10 epochs and each epoch will produce a model. System testing is carried out based on the model generated in the training process. Each model will be observed results in the form of accuracy which is calculated using a confusion matrix. The most optimal model was generated from training in epoch the 9th which resulted in an accuracy of 94.81%.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.