Joko Lianto Buliali

Generalized Regression Neural Network for predicting traffic flow

Forecasting traffic flow is a popular research topic in Intelligent Transportation System. There have been several methods used for this forecasting, such as statistical methods, Bayesian Network, Neural Network Model, Hybrid ARIMA and ANN. Generalized Regression Neural Network (GRNN) is an interesting model to be used in forecasting traffic flow, as it can predict data with dynamic change and non-linear in nature, which is generally found in traffic flow data. In this research, a GRNN model is set up to process traffic flow data, and comparing its results and the results from other predicting methods (ARIMA, Single Exponential Smoothing, and Moving Average). Leave One Out Cross Validation (LOOCV) is used in testing traffic flow data and Mean Absolute Percentage Error (MAPE) is used as the evaluation criterion in the testing. The results show that using GRNN method on the testing data can improve the accuracy of predictions by reducing the value of MAPE when three other predicting methods: ARIMA, Single Exponential Smoothing, and Moving Average.

Deteksi Jenis Kendaraan di Jalan Menggunakan OpenCV

Jenis kendaraan yang melewati suatu ruas jalan dapat diketahui secara komputasi dengan mencocokkan fitur kendaran yang terdeteksi dengan fitur kendaraan standar masing-masing jenis kendaraan . Dengan mencapture frame image yang memuat kendaraan di jalan , OpenCV dapat mencocokkan fitur kendaraan tersebut dengan fitur kendaraan standar masing-masing jenis kendaraan , sehingga jenis kendaraan pada frame image dapat diketahui .

Predictive modeling of the first year evaluation based on demographics data: Case study students of Telkom University, Indonesia

Student academic failure prediction is still interesting topics in the Educational Data Mining. One of the challenges is how to predict student academic failure as early as possible. This research focuses on predictive modeling of unsuccessful students in the first year evaluation. We propose a new concept of predictive modeling of the first year evaluation which combines 3 input data: demographics, academic and social media. The modeling can be divided into two sub modeling (normal period and extra period). In this paper, we focus on demographic data modeling (first sub-modeling) which correlated with the probability of a student to pass the first year evaluation on normal period. A Weka tool is used to get a pattern of data by using white box classifier (decision tree and rule base). Meanwhile, to solve the problem of unbalanced in our training data, we use data balancing scenario using same portion oversampling, random oversampling and SMOTE. From the testing result, we choose the best three student failure pattern of the F-Measure minor class value which obtained from “One R” and “ADTree” algorithms using Balancing scenario, the reason is because F-Measure describes the smallest error rate both FP (False Positive) and also FN (False Negative). From the best three of student failure pattern, we found that gender, selection path, study program and age are the attributes that are most correlated with the probability to pass the first year evaluation on extra period.

Metode Hibrida K-Means dan Generalized Regression Neural Network Untuk Prediksi Arus Lalu Lintas

Abstract. Traffic flow forecasting is a popular research topic in the development of Intelligent Transportation System. There have been many forecasting methods used for traffic flow forecasting, such as Generalized Regression Neural Network (GRNN) which has a fairly good accuracy. One of the GRNN’s characteristics is that the number of neurons in pattern layer increases as the number of training samples raise and this can cause overfitting problem. In this research, a hybrid method to predict traffic flow is proposed, that is K-means and GRNN algorithm. K-means method aims to solve overfitting problem in GRNN model by choosing training samples based on their similar characteristics. Leave One Out Cross Validation (LOOCV) is used to select an appropriate smoothing factor parameter at each GRNN’s model. Mean Absolute Percentage Error (MAPE) is used as the evaluation criterion in the testing process. The results show that the proposed method could improve the accuracy of predictions by reducing the value of MAPE by 0.82-3.81%. Keywords: Traffic flow forecasting, K-means, Generalized Regression Neural Network, Leave One Out Cross Validation Abstrak. Prediksi arus lalu lintas telah menjadi tren topik penelitian untuk pengembangan sistem transportasi cerdas. Telah banyak metode yang digunakan terkait prediksi arus lalu lintas, diantaranya yaitu Generalized Regression Neural Network (GRNN) yang memiliki akurasi yang cukup baik. Salah satu karakteristik GRNN adalah jumlah neuron pada pattern layer akan bertambah seiring meningkatnya jumlah data latih yang akan mengakibatkan masalah overfitting. Dalam penelitian ini diusulkan metode hibrida K-means dan GRNN untuk prediksi arus lalu lintas. Metode K-means bertujuan untuk mengatasi masalah overfitting pada model GRNN dengan memilih data latih berdasarkan kemiripan karateristiknya. Algoritma Leave One Out Cross Validation (LOOCV) digunakan untuk memilih parameter smoothing factor terbaik pada setiap model GRNN. Mean Absolute Percentage Error (MAPE) digunakan sebagai kriteria evaluasi model prediksi. Hasil menunjukkan bahwa metode yang diusulkan dapat meningkatkan akurasi prediksi dengan penurunan nilai MAPE sebesar 0 , 82-3 , 81%. Kata Kunci: Prediksi arus lalu lintas, K-means, Generalized Regression Neural Network, Leave One Out Cross Validation

Internet Connection Control based on Idle Time Using User Behavior Pattern Analysis

The increase of smartphone ability is rapidly increasing the power consumption. Many methods have been proposed to reduce smartphone power consumption. Most of these methods use the internet connection control based on the availability of the battery power level regardless of when and where a waste of energy occurs. This paper proposes a new approach to control the internet connection based on idle time using user behavior pattern analysis. User behavior patterns are used to predict idle time duration. Internet connection control performed during idle time. During idle time internet connection periodically switched on and off by a certain time interval. This method effectively reduces a waste of energy. Control of the internet connection does not interfere the user because it is implemented on idle time. Keywords : Smartphone, User Behavior, Pattern Recognition, Idle Time, Internet Connection Control

SISTEM REKOMENDASI INDEKS WEB DENGAN METODE FREQUENT TERMS BERBASIS MULTI INSTANCE LEARNING

The issues of security in mobile phone in recent days become crucial. Many privacy or secretly data is stored using unsecured protocol or sometimes without the security procedures at all. This will lead to great awareness about security in mobile phone. The effective ways to secure data are steganography and cryptography. The first one concentrate to data hiding in a certain media. In this paper, we present Chaotic Least Significant Bit Encoding (CLSBE) as a steganography method in our system design. The experiment results show that hidden messages in PNG form can be retrieved correctly. The implementation of system in emulator works well but depends on mobile phone features and environment.In computer graphics applications, to produce realistic images, a method that is often used is ray tracing. Ray tracing does not only model local illumination but also global illumination. Local illumination count ambient, diffuse and specular effects only, but global illumination also count mirroring and transparency. Local illumination count effects from the lamp(s) but global illumination count effects from other object(s) too. Objects that are usually modeled are primitive objects and mesh objects. The advantage of mesh modeling is various, interesting and real-like shape. Mesh contains many primitive objects like triangle or square (rare). A problem in mesh object modeling is long rendering time. It is because every ray must be checked with a lot of triangle of the mesh. Added by ray from other objects checking, the number of ray that traced will increase. It causes the increasing of rendering time. To solve this problem, in this research, new methods are developed to make the rendering process of mesh object faster. The new methods are angle comparison and distance comparison. These methods are used to reduce the number of ray checking. The rays predicted will not intersect with the mesh, are not checked weather the ray intersects the mesh. With angle comparison, if using small angle to compare, the rendering process will be fast. This method has disadvantage, if the shape of each triangle is big, some triangles will be corrupted. If the angle to compare is bigger, mesh corruption can be avoided but the rendering time will be longer than without comparison. With distance comparison, the rendering time is less than without comparison, and no triangle will be corrupted.Web index page is well known as page that arranges information by giving the title and short explanation about the information, where the complete information will be presented in other page. However since the amount of information become accumulate, the existence of a lot of index page exactly cause difficulty on getting information because it is possible to direct users into a mount of irrelevant information. Without a system which can help user navigation, the process of seeking the expected information is equal to a trial and error processing. In this paper, web index recommendation system is investigated which involved the activity of user on accessing the index page. This system will arrange the frequent term in index page and then implement Multi Instance Learning to give recommendation of the new index page automatically. The algorithm is citation kNN that will be adapted into fretCit kNN by implementing the minimal Hausdorff distance in measuring the distance. The experiments show that from the several test of users, the system give performance in average recommendation until 82,41% accuracy with 66,71% recall.

SISTEM BERBASIS PENGETAHUAN UNTUK KENAIKAN PANGKAT MILITER TNI AU

The issues of security in mobile phone in recent days become crucial. Many privacy or secretly data is stored using unsecured protocol or sometimes without the security procedures at all. This will lead to great awareness about security in mobile phone. The effective ways to secure data are steganography and cryptography. The first one concentrate to data hiding in a certain media. In this paper, we present Chaotic Least Significant Bit Encoding (CLSBE) as a steganography method in our system design. The experiment results show that hidden messages in PNG form can be retrieved correctly. The implementation of system in emulator works well but depends on mobile phone features and environment.In computer graphics applications, to produce realistic images, a method that is often used is ray tracing. Ray tracing does not only model local illumination but also global illumination. Local illumination count ambient, diffuse and specular effects only, but global illumination also count mirroring and transparency. Local illumination count effects from the lamp(s) but global illumination count effects from other object(s) too. Objects that are usually modeled are primitive objects and mesh objects. The advantage of mesh modeling is various, interesting and real-like shape. Mesh contains many primitive objects like triangle or square (rare). A problem in mesh object modeling is long rendering time. It is because every ray must be checked with a lot of triangle of the mesh. Added by ray from other objects checking, the number of ray that traced will increase. It causes the increasing of rendering time. To solve this problem, in this research, new methods are developed to make the rendering process of mesh object faster. The new methods are angle comparison and distance comparison. These methods are used to reduce the number of ray checking. The rays predicted will not intersect with the mesh, are not checked weather the ray intersects the mesh. With angle comparison, if using small angle to compare, the rendering process will be fast. This method has disadvantage, if the shape of each triangle is big, some triangles will be corrupted. If the angle to compare is bigger, mesh corruption can be avoided but the rendering time will be longer than without comparison. With distance comparison, the rendering time is less than without comparison, and no triangle will be corrupted.Web index page is well known as page that arranges information by giving the title and short explanation about the information, where the complete information will be presented in other page. However since the amount of information become accumulate, the existence of a lot of index page exactly cause difficulty on getting information because it is possible to direct users into a mount of irrelevant information. Without a system which can help user navigation, the process of seeking the expected information is equal to a trial and error processing. In this paper, web index recommendation system is investigated which involved the activity of user on accessing the index page. This system will arrange the frequent term in index page and then implement Multi Instance Learning to give recommendation of the new index page automatically. The algorithm is citation kNN that will be adapted into fretCit kNN by implementing the minimal Hausdorff distance in measuring the distance. The experiments show that from the several test of users, the system give performance in average recommendation until 82,41% accuracy with 66,71% recall.