Journal of Informatics And Telecommunication Engineering
Not a member yet
373 research outputs found
Sort by
Classification of the IDR-USD Exchange Rate with Multilayer Perceptron Based on Detection Rate
Get PDFArtificial neural network (ANN) is a set of units in processing a model based on the habits of human neural networks. ANN has one of its duties, namely classification with the concept of supervised learning. ANN also has various methods in performing its duties such as Multilayer perceptron (MLP). Where MLP is one of the ANN methods that can classify based on data as conceptualized in data mining. Very useful classifications and trends in the field of research because of the review of data that will generate knowledge. Nominal Exchange Rate is one of the datasets tested in this study. The exchange rate of the Indonesian Rupiah (IDR) against the United States Dollar (USD) is very necessary both in terms of stock movements and other businesses. So that, it is necessary to use classification to predict future exchange rates. In this study, the MLP method was carried out by obtaining a validation test using MAPE based on the detection rate of sebesar 0.500079879%
Comparison of C4.5 and Naïve Bayes Algorithms for Assessment of Public Complaints Services
Get PDFPublic service is one type of service provided by the government. The National Commission on Human Rights as a state institution, one of its functions is to provide services for complaints of cases of human rights violations. The purpose of this study was to find the most appropriate algorithm method by looking at the results of the accuracy and the Area Under Curve (AUC) value. The data used is data from questionnaires regarding assessments related to complaints of cases of human rights violations by the public in 2018, totaling 1750 records. The data is processed using the C4.5 algorithm and Naïve Bayes with the Rapid Miner tools. The results showed that the C4.5 Algorithm has a better accuracy of 99.49% compared to Naïve Bayes of 95.66%. The AUC value produced by the C4.5 algorithm is better at 0.998 and Naïve Bayes by 0.996. In this study, the rule generated by C4.5 will be the basis for making a questionnaire assessment application in the form of visual programming, to help provide an assessment of the satisfaction of complaint services at Komnas HAM. The system is built based on web, using PHP framework, database using MySQL and editor tools using notepad ++
An Analysis of Contract Employee Performance Assessment Using Machine Learning
Get PDFSumber daya manusia menjadi salah satu perhatian utama bagi pengelola semua jenis bisnis baik perusahaan negara maupun perusahaan swasta. Organisasi usaha sangat tertarik dalam menyusun rencana dan eksekusi untuk mempertahankan karyawan kontrak yang berkinerja baik. Dilain pihak banyak nya perpanjangan karyawan kontrak berdasarkan hubungan kedekatan keluarga maupuan kedekatan kerabat. Untuk mendapatkan pekerja yang berkinerja baik, dibutuhkan aplikasi assesment kinerja karyawan kontrak. Dengan menggunakan pendekatan algoritma machine learning seperti Random Forest, Decision Tree, K-Nearest Neighbors, Naïve Bayes, dan Logistic Regression. Penelitian ini mencoba mencari algoritma yang tepat untuk aplikasi assesment kinerja karyawan kontrak dengan nilai input skill, attidude, responsibility, absent dan membandingkan nilai akurasi, presisi, recall dan F1 score. Berdasarkan penelitian, model algoritma Random Forest yang terbaik digunakan dalam mengklasifikasi dan prediksi assement kinerja karyawan kontrak dengan mencapai nilai akurasi 90.62 %, presisi 72.22 %, recall 75% dan nilai F1 score 71.11
Initial Centroid Optimization of K-Means Algorithm Using Cosine Similarity
Get PDFClustering salah satu metode yang sering digunakan di berbagai bidang yang melakukan analisis data, termasuk penggalian data, pengambilan dokumen, segmentasi gambar, dan klasifikasi pola. Adapaun tujuan dari metode tersebut adalah untuk mengelompokkan data ke dalam suatu cluster sehingga kesamaan antara anggota data dalam suatu data informasi yang telah di-cluster yang sama adalah maksimal, di sisi lain untuk kesamaan di antara anggota data yang lain berbeda cluster minimal. Ada beberapa pendekatan metode untuk mengurangi kesalahan pada saat centroid awal yang dipilih selama proses pengelompokan berlangsung. Disini data yang digunakan adalah data acak yang dibuat secara manual yatu 30 data dan 5 atribut, sehingga diperoleh hasil akurasi clustering dalam centroid dengan menggunakan metode K-Means memiliki signifikan 86.67%, sedangkan menggunakan K-Means dengan cosine similarity tidak jauh berbeda yaitu sebesar 89.7%, maka dari itu hasilnya cukup baik
Analysis of Air Pollution Levels in DKI Jakarta Province Using the Mamdani Fuzzy Inference System Method
Get PDFThis study aims to measure the level of air pollution determined by pollutant gases contained in the air. Pollutants that measure air pollution are PM10 (Special Material), SO2 (Sulfur), NO2 (Nitrogen Oxide), CO (Carbon Monoxide, O3 (Ozone), and NO2 (Nitrogen Oxide), which are related to vehicle use and, according to the choice this pollutant threshold, we will discuss the level of air pollution with the fuzzy mamdani inference method. The results of the pollutant threshold study will then be applied to the rules / rules that are applied using the if-then rules and then the input variables are arranged using weighted averages, variable averages weighted will be determined higher into three levels: low, medium and high.Keywords Decision Tree, Feature Selection, Optimization of Lecturer Assistant Performance, Particle Swarm Optimization
Face Recognition based Feature Extraction using Principal Component Analysis (PCA)
Get PDFThe human face is an entity that has semantic features. Face detection is the first step before face recognition. Face recognition technique is an identification process based on facial features. One feature extraction approach for facial recognition techniques is the Principal Component Analysis (PCA) method. The PCA method is used to simplify facial features and characteristics in order to obtain proportions that are able to represent the characteristics of the original face. The purpose of this research is to construct facial patterns stored in a digital image database. The process of pattern construction and face recognition starts from objects in the form of face images, side detection, pattern construction until it can determine the similarity of face patterns to proceed as face recognition. In this research, a program has been designed to test some samples of face data stored in a digital image database so that it can provide a similarity in the face patterns being observed and its introduction using PC
An Analysis of Slot Dimension Changing in Dual band Rectangular Patch Microstrip Antenna with Proximity Coupled Feed
Get PDFIn this paper, the characteristics of dual band rectangular patch microstrip antenna using proximity couple feed are studied. It can be used for a wireless device that works on multiband frequency. The addition of slot and proximity feed used in order to obtain larger bandwidth and multiple frequency. Microstrip antenna is designed and simulated using software also used to analyze by changing the variable of microstrip slot’s dimension. The parameters are tested in this study include Voltage standing wave ratio (VSWR), return loss, gain, bandwidth and radiation patterns. From the simulation results, the best value of return loss antenna is -23,29 dB at 2,4 GHz with a slot width of 1 mm and 0,085 GHz bandwidth. At 3,7 GHz, the best value of return loss antenna is -23dB with a slot width of 2 mm and 0,12 GHz bandwidth. Afterwards, the best VSWR obtained on dual band microstrip antennas with proximity coupled feed is 1,14 and 5.53 dBi gain.Keywords: slot, bandwidth, proximity, return loss, gain
Sentiment Analysis on Corona Virus Pandemic Using Machine Learning Algorithm
Get PDFCorona virus outbreaks that occur in almost all countries in the world have an impact not only in the health sector, but also in other sectors such as tourism, finance, transportation, etc. This raises a variety of sentiments from the public with the emergence of corona virus as a trending topic on Twitter social media. Twitter was chosen by the public because it can disseminate information in real time and can see market reactions quickly. This research uses "tweet" data or public tweet related to "Corona Virus" to see how the sentiment polarity arises. Text mining techniques and three machine learning classification algorithms are used, including Naive Bayes, Support Vector Machine (SVM), K-Nearest Neighbor (K-NN) to build a tweet classification model of sentiments whether they have positive, negative, or neutral polarity. The highest test results are generated by the Support Vector Machine (SVM) algorithm with an accuracy value of 76.21%, a precision value of 78.04%, and a recall value of 71.42%.Keywords: Machine Learning, Corona Virus, Twitter, Sentiment Analysis
Comparative Compression of Wavelet Haar Transformation with Discrete Wavelet Transform on Colored Image Compression
Get PDFIn this research, the algorithm used to compress images is using the haar wavelet transformation method and the discrete wavelet transform algorithm. The image compression based on Wavelet Wavelet transform uses a calculation system with decomposition with row direction and decomposition with column direction. While discrete wavelet transform-based image compression, the size of the compressed image produced will be more optimal because some information that is not so useful, not so felt, and not so seen by humans will be eliminated so that humans still assume that the data can still be used even though it is compressed. The data used are data taken directly, so the test results are obtained that digital image compression based on Wavelet Wavelet Transformation gets a compression ratio of 41%, while the discrete wavelet transform reaches 29.5%. Based on research problems regarding the efficiency of storage media, it can be concluded that the right algorithm to choose is the Haar Wavelet transformation algorithm. To improve compression results it is recommended to use wavelet transforms other than haar, such as daubechies, symlets, and so on
Algorithm Implementation Of Interest Buy Apriori Data On Consumer Retail Sales In Industry
Get PDFThis data processing has the aim to increase the company's turnover, because by being aware of how the interest in buying goods works, the company can buy products other than the main products that it buys. In increasing company revenue can be done using the Data Mining process, one of which uses a priori algorithm and association techniques. With this a priori algorithm found association technique which later can be used as a pattern of purchasing goods by consumers, this study uses a data repository of 958 data consisting of 45 transactions. From the results obtained goods with the name Paper Chain Kit 50's Christmas is a product that is often bought by consumers and it is known that the most frequent combination patterns are the Retro Spot Paper Chain Kit and the Paper Chain Kit 50's Christmas. So that with known buying patterns, the company manager can predict future market needs, and can calculate the stock of goods that must be reproduced, and goods whose stock must be reduced, and also with the results of the association the manager can manage the layout of the product to be better.Keywords: Apriori Algorithm, Sales Data, Retail