Jurnal Teknologi dan Sistem Komputer
Not a member yet
364 research outputs found
Sort by
Comparison of the histogram of oriented gradient, GLCM, and shape feature extraction methods for breast cancer classification using SVM
Full text linkBreast cancer originates from the ducts or lobules of the breast and is the second leading cause of death after cervical cancer. Therefore, early breast cancer screening is required, one of which is mammography. Mammography images can be automatically identified using Computer-Aided Diagnosis by leveraging machine learning classifications. This study analyzes the Support Vector Machine (SVM) in classifying breast cancer. It compares the performance of three features extraction methods used in SVM, namely Histogram of Oriented Gradient (HOG), GLCM, and shape feature extraction. The dataset consists of 320 mammogram image data from MIAS containing 203 normal images and 117 abnormal images. Each extraction method used three kernels, namely Linear, Gaussian, and Polynomial. The shape feature extraction-SVM using Linear kernel shows the best performance with an accuracy of 98.44 %, sensitivity of 100 %, and specificity of 97.50 %
Implementing a non-local means method to CTA data of aortic dissection
Full text linkIt is necessary to conserve important information, like edges, details, and textures, in CT aortic dissection images, as this helps the radiologist examine and diagnose the disease. Hence, a less noisy image is required to support medical experts in performing better diagnoses. In this work, the non-local means (NLM) method is conducted to minimize the noise in CT images of aortic dissection patients as a preprocessing step to produce accurate aortic segmentation results. The method is implemented in an existing segmentation system using six different kernel functions, and the evaluation is done by assessing DSC, precision, and recall of segmentation results. Furthermore, the visual quality of denoised images is also taken into account to be determined. Besides, a comparative analysis between NLM and other denoising methods is done in this experiment. The results showed that NLM yields encouraging segmentation results, even though the visualization of denoised images is unacceptable. Applying the NLM algorithm with the flat function provides the highest DSC, precision, and recall values of 0.937101, 0.954835, and 0.920517 consecutively
Enhanced image security using residue number system and new Arnold transform
Full text linkThis paper aims to improve the image scrambling and encryption effect in traditional two-dimensional discrete Arnold transform by introducing a new Residue number system (RNS) with three moduli and the New Arnold Transform. The study focuses on improving the classical discrete Arnold transform with quasi-affine properties, applying image scrambling and encryption research. The design of the method is explicit to three moduli set {2n, 2n+1+1, 2n+1-1}. These moduli set includes equalized and shapely moduli leading to the effective execution of the residue to binary converter. The study employs an arithmetic residue to the binary converter and an improved Arnold transformation algorithm. The encryption process uses MATLAB to accept a digital image input and subsequently convert the image into an RNS representation. The images are connected as a group. The resulting encrypted image uses the Arnold transformation algorithm. The encrypted image is used as input at decryption using the anti-Arnold (Reverse Arnold) transformation algorithm to convert the picture to the original RNS (original pixel value). Then the RNS was used to retransform the original RNS to its binary form. Security analysis tests, like histogram analysis, keyspace, key sensitivity, and correlation coefficient analysis, were administered on the encrypted image. Results show that the hybrid system can use the improved Arnold transform algorithm with better security and no constraint on image width and size
Comparison analysis of Euclidean and Gower distance measures on k-medoids cluster
Full text linkKlastering k-medoids menggunakan metode jarak untuk mencari dan mengelompokkan data yang memiliki kesamaan dan ketidaksamaan. Penentuan metode pengukuran jarak adalah hal yang penting karena mempengaruhi performa hasil klaster k-medoids. Beberapa kajian menyatakan bahwa metode Euclidean dan Gower bisa digunakan sebagai metode pengukuran pada klastering dengan data numerik. Penelitian ini bertujuan untuk melakukan perbandingan performa hasil klastering k-medoids pada dataset numerik menggunakan metode Euclidean dan Gower. Penelitian ini menggunakan tujuh dataset numerik dan evaluasi hasil klastering menggunakan nilai Silhouette, Dunn, dan Connectivity. Metode jarak Euclidean unggul pada dua nilai evaluasi Silhouette dan Connectivity yang menunjukkan bahwa Euclidean memiliki struktur pengelompokan data yang baik, sedangkan Gower unggul pada satu nilai evaluasi Dunn yang menunjukkan Gower memiliki pemisah antar klaster yang baik dibanding Euclidean. Penelitian ini menunjukkan bahwa metode Euclidean lebih unggul daripada metode Gower pada penerapan algoritma k-medoids dengan dataset bertipe numerik.K-medoids clustering uses distance measurement to find and classify data that have similarities and inequalities. The distance measurement method selection can affect the clustering performance for a dataset. Several studies use the Euclidean and Gower distance as measurement methods in numerical data clustering. This study aims to compare the performance of the k-medoids clustering on a numerical dataset using the Euclidean and Gower distance. This study used seven numerical datasets and Silhouette, Dunn, and Connectivity indexes in the clustering evaluation. The Euclidean distance is superior in two values of Silhouette and Connectivity indexes so that Euclidean has a good data grouping structure, while the Gower is superior in Dunn index showing that the Gower has better cluster separation compared to Euclidean. This study shows that the Euclidean distance is superior to the Gower in applying the k-medoids algorithm with a numeric dataset
Prediction of hotel bookings cancellation using hyperparameter optimization on Random Forest algorithm
Full text linkPembatalan pemesanan hotel oleh pelanggan sangat mempengaruhi pengambilan keputusan manajerial hotel. Untuk meminimalkan kerugian akibat masalah ini, pihak pengelola hotel membuat kebijakan yang cukup kaku yang dapat merusak reputasi dan kinerja bisnis. Oleh karena itu, penelitian ini berfokus pada penyelesaian masalah tersebut menggunakan algoritme pembelajaran mesin. Untuk mendapatkan performa model terbaik, optimalisasi hiperparameter diterapkan pada algoritme random forest. Hal tersebut bertujuan untuk mendapatkan kombinasi parameter model yang terbaik dalam memprediksi pembatalan pemesanan hotel. Model yang diusulkan terbukti memiliki kinerja terbaik dengan hasil akurasi tertinggi 87 %. Hasil dari penelitian ini dapat digunakan sebagai komponen model dalam sistem pengambilan keputusan manajerial hotel terkait pembatalan pemesanan di masa mendatang.Cancellation of hotel bookings by customers greatly influences hotel managerial decision making. To minimize losses by this problem, the hotel management made a fairly rigid policy that could damage the reputation and business performance. Therefore, this study focuses on solving these problems using machine learning algorithms. To get the best model performance, hyperparameter optimization is applied to the random forest algorithm. It aims to obtain the best combination of model parameters in predicting hotel booking cancellations. The proposed model is proven to have the best performance with the highest accuracy results of 87 %. This study's results can be used as a model component in hotel managerial decision-making systems related to future bookings' cancellation
Erratum: Optimasi nilai k dan parameter lag algoritme k-nearest neighbor pada prediksi tingkat hunian hotel
Full text linkThis correct the article "Optimasi nilai k dan parameter lag algoritme k-nearest neighbor pada prediksi tingkat hunian hotel (Optimization of k value and lag parameter of k-nearest neighbor algorithm on the prediction of hotel occupancy rates)" in vol. 8, no. 3, pp. 246-254, Jul. 2020; https://doi.org/10.14710/jtsiskom.2020.13648In the original published article, the placement of Figure 8 and Figure 9 less appropriate, which causes the manuscript hard to read. In addition, Table 2 through Table 6 need to be repositioned. These placing errors have been corrected online.The publisher apologizes for these errors. Perbaikan ini dilakukan terhadap artikel “Optimasi nilai k dan parameter lag algoritme k-nearest neighbor pada prediksi tingkat hunian hotel (Optimization of k value and lag parameter of k-nearest neighbor algorithm on the prediction of hotel occupancy rates)" di vol. 8, no. 3, pp. 246-254, Jul. 2020; https://doi.org/10.14710/jtsiskom.2020.13648.Dalam versi original artikel terbit, penempatan Gambar 8 dan 9 kurang sesuai sehingga menyebabkan naskah sulit untuk dibaca. Selain itu, Tabel 2 dan Tabel 6 perlu untuk dipindahkan untuk menyesuaikan perbaikan ini.Penerbit mohon maaf atas kekurangan ini. Perbaikan telah dilakukan dan revisi artikel telah diunggah
Pengenalan rambu lalu lintas menggunakan convolutional neural networks
Full text linkTraffic sign recognition (TSR) can be used to recognize traffic signs by utilizing image processing. This paper presents traffic sign recognition in Indonesia using convolutional neural networks (CNN). The overall image dataset used is 2050 images of traffic signs, consisting of 10 kinds of signs. The CNN layer used in this study consists of one convolution layer, one pooling layer using maxpool operation, and one fully connected layer. The training algorithm used is stochastic gradient descent (SGD). At the training stage, using 1750 training images, 48 filters, and a learning rate of 0.005, the recognition results in 0.005 of loss and 100 % of accuracy. At the testing stage using 300 test images, the system recognizes the signs with 0.107 of loss and 97.33 % of accuracy.Traffic sign recognition (TSR) digunakan mengenali rambu lalu lintas dengan memanfaatkan pengolahan citra. Artikel ini menyajikan pengenalan rambu lalu lintas di Indonesia menggunakan convolutional neural networks (CNN). Dataset citra yang digunakan secara keseluruhan adalah 2050 citra rambu lalu lintas, yang terdiri dari 10 macam rambu. Lapisan CNN yang digunakan terdiri dari satu lapisan konvolusi, satu lapisan pooling menggunakan operasi maxpool, dan satu lapisan fully-connected. Algoritme pelatihan yang digunakan adalah Stochastic Gradient Descent (SGD). Pada tahap pelatihan dengan menggunakan 1750 data citra latih, 48 filter dan laju pelatihan 0,005, dihasilkan galat 0,005 dan akurasi 100 %. Pada tahap pengujian menggunakan 300 data citra uji, sistem dapat mengenali rambu lalu lintas dengan galat 0,107 dan akurasi mencapai 97,33 %
Analisis penerapan tapis Wiener pada segmentasi pola fluktuasi spektral
Full text linkThe Wiener filter is an adaptive filter which able to produce the desired estimates. Besides, this filter can also suppress noise in digital signal processing. This study aims to segment the fluctuation pattern, which results from data acquisition from a capacitive sensor with the object H2O. The fluctuation pattern to be processed is the High Fluctuation (HF) pattern by dividing the pattern into several segments according to the input frequency. It aims to see in more detail and clearly the state of each segmentation of the pattern. The results will show noise attenuation and suppression after filtering with a Wiener filter. The Signal to Noise Ratio (SNR) value will also be analyzed, which shows that the signal quality is getting better after applying the Wiener filter. Then, the analysis of the Mean Square Error (MSE) results can provide more consistent results with a smaller average error.Tapis Wiener merupakan suatu tapis adaptif yang dapat digunakan untuk menghasilkan perkiraan yang diinginkan. Selain itu, tapis ini juga dapat menekan derau pada pengolahan sinyal digital. Kajian melakukan segmentasi terhadap pola fluktuasi yang merupakan hasil akuisisi data dari sebuah sensor kapasitif dengan objeknya H2O. Pola fluktuasi yang diolah adalah pola fluktuasi tinggi (HF, High Fluctuation) dengan cara membagi pola tersebut ke dalam beberapa segmen sesuai dengan frekuensi masukan. Hal ini bertujuan untuk dapat melihat lebih detil dan jelas keadaan setiap segmentasi dari pola tersebut. Hasilnya menunjukkan peredaman dan penekanan derau setelah ditapis dengan tapis Wiener. Selain itu, nilai SNR juga dianalisis dan menunjukkan bahwa kualitas sinyal semakin baik sesudah penerapan tapis Wiener. Analisis hasil nilai MSE mampu memberikan hasil yang lebih konsisten dengan rata-rata kesalahan yang lebih kecil
Attitude stabilization control for quadrotor using self-tuning fuzzy-PD
Full text linkPenelitian ini bertujuan mengembangkan sistem pengendalian attitude quadrotor dalam mempertahankan posisi dan keseimbangan terhadap gangguan pada saat terbang (hovering). Quadrotor membutuhkan sistem kendali non-linear yang handal dan cepat pada kondisi hovering untuk melakukan respons terhadap manuver yang tinggi pada sistem dengan enam derajat kebebasan, sehingga penelitian ini berfokus untuk merancang dan menguji sistem kendali Self-Tuning Fuzzy-PD untuk kendali attitude quadrotor. Sistem kendali quadrotor dirancang menggunakan data input dari INS (Inertial Navigation System). Selanjutnya pengendalian attitude quadrotor dilakukan dengan meneruskan output sinyal PWM hasil komputasi ke flight controller APM 2.6. Berdasarkan hasil pengujian, diperoleh rata-rata galat absolut yang cukup kecil untuk sudut roll, pitch, dan yaw secara berurutan sebesar 2,079o, 2,266o, dan 1,528o, sedangkan galat absolut maksimalnya sebesar 6,314o, 6,722o, dan 3,82o.This research aims to develop a quadrotor control system for maintaining its position and balance from disturbance while hovering. A fast and reliable control technique is required to respond to high maneuverability and high non-linearity of six degrees of freedom system. Hence, this research focuses on designing a Self-Tuning Fuzzy-PD control system for quadrotor’s attitude. The designed control system utilizes input data from the Inertial Navigation System (INS). Then the quadrotor’s attitude is controlled by passing the PWM signal to the flight controller APM 2.6. The result shows that the average absolute error for the roll, pitch, and yaw angles are relatively small, as mentioned consecutively 2.079o, 2.266o, and 1.528o, while the maximum absolute errors are 6.314o, 6.722o, and 3.82o
Rice consumption prediction using linear regression method for smart rice box system
Full text linkKotak beras cerdas dapat menerapkan Internet of Things (IoT), namun belum menerapkan prediksi kapan beras habis yang menunjukkan jumlah konsumsi beras. Penelitian ini mengkaji penerapan regresi linier untuk prediksi waktu habis beras pada kotak beras cerdas berbasis IoT dan menganalisis kinerjanya. Prediksi regresi linier dibuat berdasarkan dataset yang diperoleh dengan mengukur kotak beras cerdas yang telah dipasang sensor berat load cell dan modul Hx711 dengan RMSE pembacaan berkisar antara 56 hingga 170 gram. Prediksi waktu habis beras dengan metode regresi linear mempunyai nilai MSE sebesar 0,2588 dengan jendela prediksi 43 hari. Nilai R-quared kurang dari 1 diperoleh dengan ambang prediksi 24 hari.Currently, the smart rice box has applied the Internet of Things (IoT) but without prediction of rice runs out which shows the amount of rice consumption. This study applies linear regression to predict the rice runs out in an IoT-based smart rice box and analyzes its performance. The prediction used the dataset obtained by measuring a smart rice box equipped with a load cell weight sensor and Hx711 module. The weight sensor accuracy was an RMSE of between 56 and 170 grams. The linear regression method applied to the smart rice box to predict rice running out has an MSE value of 0.2588 with a prediction window of 43 days. An R-squared value of less than one is obtained with a predictive threshold of 24 days