PENS Repository
Not a member yet
1406 research outputs found
Sort by
Intrusion Detection with Classification and On-Line Clustering Using Reinforcement Learning
No full textToday, information technology is growing rapidly,all information can be obtained much easier. It raises some new problems, one of them is unauthorized access to the system. We need a reliable network security system that is resistant to a variety of attacks against the system. Therefore, Intrusion Detection System (IDS) required to overcome the problems of intrusions. Many researches have been done on intrusion detection using classification methods. Classification methodshave high precision, but it takes efforts to determine an appropriate classification model to the classification problem. In this paper, we propose a new reinforced approach to detect intrusion with On-line Clustering using Reinforcement Learning. Reinforcement Learning is a new paradigm in machine learning which involves interaction with the environment.It works with reward and punishment mechanism to achieve solution. We apply the Reinforcement Learning to the intrusion detection problem with considering competitive learning using Pursuit Reinforcement Competitive Learning (PRCL). Based on the experimental result, PRCL can detect intrusions in real time with high accuracy (99.816% for DoS, 95.015% for Probe, 94.731% for R2L and 99.373% for U2R) and high speed (44 ms). The proposed approach can help network administrators to detect intrusion, so the computer network security systembecome reliable
Impression Generation of Indonesian Cultural Paintings for Mobile Application with Culture Dependent Color-Impression Metric Creation Contents
No full textPainting is one of complex image reflecting observations and feelings of the artist to the environment. This condition extends the need of painting impression generation system since common people with lack of art experience would have difficulties to interpret the painting. From this point of view we presents a new model to provide representative impressions of paintings by providing a color-impression metric taken from public survey and implement it for mobile application. The new model provides analytical functions to generate the representative impression of the image query. The functions consist of two main section: (1) The cultural-dependent color-impression metric creation which consist of conducting survey, applying normalized 3D color vector quantization to image dataset, generating image-impression metric, and generating color- impression metric; and (2) Impression generation of image query which consist of applying normalized 3D color vector quantization to image query and measuring the similarity between image query and color-impression metric. To perform our proposed impression generation system, we examine our system with Indonesian cultural image dataset and 5 different mobile devices. Our proposed system performs main color impression precision result with average precision of more than 60%. Brightness intensity and zooming affects the retrieved impressions. Rotating captures of an image generate the same retrieved impressions. The system also performs average response time vary in range 41263 to 117434 milliseconds from all devices
Centronit: Initial Centroid Designation Algorithm for K-Means Clustering
No full textClustering performance of the K-means highly depends on the correctness of initial centroids. Usually initial centroids for the K- means clustering are determined randomly so that the determined initial centers may cause to reach the nearest local minima, not the global optimum. In this paper, we propose an algorithm, called as Centronit, for designation of initial centroid optimization of K-means clustering. The proposed algorithm is based on the calculation of the average distance of the nearest data inside region of the minimum distance. The initial centroids can be designated by the lowest average distance of each data. The minimum distance is set by calculating the average distance between the data. This method is also robust from outliers of data. The experimental results show effectiveness of the proposed method to improve the clustering results with the K-means clustering
Earthquake Density Measurement using Automatic Clustering
No full textThe government and earthquake associations have recorded the seismic data in spatial-temporal usually used to measure the earthquake intensity, but such information has not been processed to obtain the earthquake density. This condition makes it difficult to map all the risk, so it creates the lack of participatory development to the earthquake areas that have a high density. This research proposes a new approach for measuring the density of earthquake and performing automatic clustering with valley tracing method to detect the number of group spatial regions automatically from analysis of cluster moving variances. The density is obtained with a new approach that involves the area and the amount of data on clusters. This system follows the steps: (1) display the spatial-temporal earthquake dataset into a map, (2) create vector space data consist of temporal, spatial, and magnitude, (3) detect number of cluster with valley tracing method on automatic clustering, (4) calculate the earthquake density, and (5) display special temporal visualization with the cluster density measurement result. To perform the proposed idea, this system is examined with an experimental study for a series of quake about Indonesia and Japan during the last 50 years from ANSS catalog
Batik Image Search System with Extracted Combination of Color and Shape Features
No full textBatik is a culture-dependent technique and symbolism surrounding hand-dyed cotton and silk garments from Indonesia It an cultural art that has a long history of acculturation, with diverse patterns influenced by a variety of cultures. Each region in Indonesian has specific colors and shapes reflecting the identity of the region. Due to many diverse Batik' patterns and colors, it is difficult to retrive the Batik images from both the color and pattern. In this paper we propose a new system for Batik image search with providing an analytical function for feature extraction by involving color and shape features and combining the extracted features. We use 3D-Vector Quantization for color feature extraction. It can uniformly represent the distribution of image colors and reduce complexity of the colors. Beside the color features, we also use shape features of the Batik by applying Hu's moment. The extracted shape features from the Hu's moment consists of orthogonal moment invariants those can be used for scale, position, and rotation invariant pattern identification for the Batik patterns. Finally, the distribution of color moments are used to aggregate the extracted color and shape features. For experimental study, we apply our proposed system to 210 Batik image dataset from 3 common Batik kinds of pattern: Kawung, Parang and Mega Mendung. The system performs the easy-to-use application for the users in which they may easily search the Batik images with involving both the color and shape features
Environmental Change Detection and Visualization by Differential Computing for Satellite Images with 5D World Map System
No full textThis paper presents a differential computing system for satellite images called SIDE, which automatically extracts areas, boundaries and lines on earth from any type of general images using image processing technology. Unlike other general GIS software, SIDE can process any type of images without geo information (lat, long) as input and convert to geo information with template. The feature of SIDE is that it enables users to extract the differences between images using color information of areas on earth such as forest, ice, ocean, river, soil and desert, and compare the differences in time-series analysis. Also, SIDE provides a function to modify the position, size and angle of detected area, border and lines, and adjust the positions to geo information to overlap output images onto the base-map such as Google Maps or ArcGIS maps. By using this system with a multimedia sharing system called 5D World Map System, users can analyze the impacts of environmental change to the other phenomena by global viewpoints with other related multimedia
Intrusion Detection with On-line Clustering Using Reinforcement Learning
No full textToday, information technology is growing rapidly, we can obtain all the information much easier. Almost all the important information can be accessed by the users. These conditions raise some new problems, one of them is unauthorized access to the system. We need a reliable network security system that is resistant to a variety of attacks against the system. Therefore, Intrusion Detection System (IDS) required to overcome the problems of intrusions. Many researches have been done on intrusion detection using classification methods. Classification method has high precision, but to get a high precision required a determination of the proper classification model. In this paper, we propose a new approach to detect intrusion with On-line Clustering using Reinforcement Learning. Based on the experimental result, our proposed technique can detect intrusions with high accuracy (99.996% for DoS, 99.939% for Probe, 99.865% for R2L and 99.948% for U2R) and high speed (65 ms)
Automatic Representative News Generation Using On-Line Clustering
No full textThe increasing number of online news provider has produced large volume of news every day. The large volume can bring drawback in consuming information efficiently because some news contain similar contents but they have different titles that may appear. This paper presents a new system for automatically generating representative news using on-line clustering. The system allows the clustering to be dynamic with the features of centroid update and new cluster creation. Text mining is implemented to extract the news contents. The representative news is obtained from the closest distance to each centroid that calculated using Euclidean distance. For experimental study, we implement our system to 460 news in Bahasa Indonesia. The experiment performed 70.9% of precision ratio. The error is mainly caused by imprecise results from keyword extraction that generates only one or two keywords for an article. The distribution of centroid’s keywords also affects the clustering results
Optimasi Pergerakan Robot Planar 3 Sendi pada Robot Penggenggam dengan menggunakan Metode Pemrograman Genetika
Full text linkPengembangan robot yang mengadopsi sistem pergerakan bagian tubuh manusia semakin pesat saat ini. Pada paper ini dijelaskan tentang pengembangan robot penggenggam (gripper) yang meniru prinsip jari telunjuk dan ibu jari manusia. Ibu jari memiliki 2 sendi sedangkan jari telunjuk memiliki 3 sendi. Jari-jari ini digerakkan dengan menggunakan motor DC yang terhubung langsung pada masing-masing sendi. Dengan menggunakan persamaan kinematika terbalik, maka sudut pergerakan masing-masing sendi dapat ditentukan untuk mengarahkan end-effector pada posisi tertentu. Permasalah yang muncul adalah akan dihasilkan beberapa nilai sudut pada sendi-sendi jari telunjuk yang memiliki 3 sendi. Pada paper ini akan digunakan metode pemrograman genetika untuk menentukan nilai sudut yang paling optimal dari masing-masing sendi pada jari telunjuk. Kriteria optimal adalah total sudut end-effector,(theta_e), terkecil untuk posisi koordinat end-effector yang diinginkan. Dengan pemrograman genetika diperoleh persamaan untuk mendapatkan theta_e paling optimal pada posisi koordinat end-effector yang diinginkan
SINKRONISASI KECEPATAN PUTARAN ELEKTRONIK SOFT SWITCH GEAR BERBASIS PID KONTROL PADA MOBIL HIBRIDA
Full text linkPada sistem penggerak mobil hibrida dalam penelitian digunakan penggerak motor elektrik dengan motor induksi tiga fasa dan penggerak mesin pembakaran dalam berbahan bakar minyak (BBM) sebagai penggerak cadangan. Untuk perpindahan kedua penggerak tersebut diperlukan sistem perpindahan transmisi penggerak yang lembut pada saat transisi perpindahan transmisi kedua penggerak tersebut. Untuk itu diperlukan teknik kendali pada elektronik switch gear pada transmisi penggerak mobil. Teknik kendali yang digunakan dengan kontrol PID untuk mensinkronkan kedua kecepatan putaran penggerak, sehingga pada saat perpindahan kedua transmisi dihasilkan switch gear yang soft. Langkah yang dilakukan adalah dengan pengaturan kecepatan putaran mesin BBM dan putaran kecepatan motor elektrik. Mesin BBM sebagai Master dan motor elektrik sebagai Slave karena pengaturan motor elektrik lebih mudah dari pada pengaturan mesin BBM, sehingga yang menyesuaikan kecepatan transmisi adalah motor elektrik. Pada saat akan dilakukan perpindahan transmisi kecepatan motor akan menyesuakan kecepatan mesin BBM dan apabila sudah dicapai sinkronisasi dengan teknik kendali PID maka dilanjutkan dengan perpindahan transmisi. Penggerak utama mobil hibrida ini adalah motor elektrik, apabila terjadi kelemahan daya pada baterai, dilakukan pemindahan transmisi ke penggerak BBM dan secara bersamaan dilakukan pengisian baterai sampai penuh. Apabila pengisian baterai sudah penuh maka dilakukan perpindahan transmisi ke penggerak motor elektrik. Hasil simulasi menunjukkan bahwa kendali PID sesuai untuk diterapkan pada mobil hibrida ini karena mempunyai respon yang baik dan bekerja secara autotuning pada perubahan kecepatan mobil.
Kata Kunci: motor induksi tiga fasa, kontrol PID, elektronik switch gear, autotuning