International Journal of Advances in Intelligent Informatics
Not a member yet
235 research outputs found
Sort by
Optimization of data resampling through GA for the classification of imbalanced datasets
Full text linkClassification of imbalanced datasets is a critical problem in numerous contexts. In these applications, standard methods are not able to satisfactorily detect rare patterns due to multiple factors that bias the classifiers toward the frequent class. This paper overview a novel family of methods for the resampling of an imbalanced dataset in order to maximize the performance of arbitrary data-driven classifiers. The presented approaches exploit genetic algorithms (GA) for the optimization of the data selection process according to a set of criteria that assess each candidate sample suitability. A comparison among the presented techniques on a set of industrial and literature datasets put into evidence the validity of this family of approaches, which is able not only to improve the performance of a standard classifier but also to determine the optimal resampling rate automatically. Future activities for the improvement of the proposed approach will include the development of new criteria for the assessment of sample suitability
Reversible data hiding method by extending reduced difference expansion
Full text linkTo keep hiding secret data in multimedia files, such as video, audio, and image considers essential for information security. Image, for instance, as the media aids data insertion securely. The use of insertion technique must ensure a reliable process on retaining data quality and capacity. However, a trade-off between the resulted image quality and the embedded payload capacity after the embedding process often occurs. Therefore, this research aims at extending the existing method of integrating confidential messages using the Reduced Difference Expansion (RDE), transform into a medical image by changing the base point, block size, and recalculating of difference. The results display that the proposed method enhances the quality of the stego image and capacity of the hidden message
Optimized biometric system based iris-signature for human identification
Full text linkThis research aimed at comparing iris-signature techniques, namely the Sequential Technique (ST) and the Standard Deviation Technique (SDT). Both techniques were measured by Backpropagation (BP), Probabilistic, Radial basis function (RBF), and Euclidian distance (ED) classifiers. A biometric system-based iris is developed to identify 30 of CASIA-v1 and 10 subjects from the Real-iris datasets. Then, the proposed unimodal system uses Fourier descriptors to extract the iris features and represent them as an iris-signature graph. The 150 values of input machine vector were optimized to include only high-frequency coefficients of the iris-signature, then the two optimization techniques are applied and compared. The first optimization (ST) selects sequentially new feature values with different lengths from the enrichment graph region that has rapid frequency changes. The second technique (SDT) chooses the high variance coefficients as a new feature of vectors based on the standard deviation formula. The results show that SDT achieved better recognition performance with the lowest vector-lengths, while Probabilistic and BP have the best accuracy
Nonstandard optimal control problem: case study in an economical application of royalty problem
Full text linkThis paper's focal point is on the nonstandard Optimal Control (OC) problem. In this matter, the value of the final state variable, y(T) is said to be unknown. Moreover, the Lagrangian integrand in the function is in the form of a piecewise constant integrand function of the unknown state value y(T). In addition, the Lagrangian integrand depends on the y(T) value. Thus, this case is considered as the nonstandard OC problem where the problem cannot be resolved by using Pontryagin’s Minimum Principle along with the normal boundary conditions at the final time in the classical setting. Furthermore, the free final state value, y(T) in the nonstandard OC problem yields a necessary boundary condition of final costate value, p(T) which is not equal to zero. Therefore, the new necessary condition of final state value, y(T) should be equal to a certain continuous integral function of y(T)=z since the integrand is a component of y(T). In this study, the 3-stage piecewise constant integrand system will be approximated by utilizing the continuous approximation of the hyperbolic tangent (tanh) procedure. This paper presents the solution by using the computer software of C++ programming and AMPL program language. The Two-Point Boundary Value Problem will be solved by applying the indirect method which will involve the shooting method where it is a combination of the Newton and the minimization algorithm (Golden Section Search and Brent methods). Finally, the results will be compared with the direct methods (Euler, Runge-Kutta, Trapezoidal and Hermite-Simpson approximations) as a validation process
Medoid-based shadow value validation and visualization
Full text linkA silhouette index is a well-known measure of an internal criteria validation for the clustering algorithm results. While it is a medoid-based validation index, a centroid-based validation index that is called a centroid-based shadow value (CSV) has been developed. Although both are similar, the CSV has an additional unique property where an image of a 2-dimensional neighborhood graph is possible. A new internal validation index is proposed in this article in order to create a medoid-based validation that has an ability to visualize the results in a 2-dimensional plot. The proposed index behaves similarly to the silhouette index and produces a network visualization, which is comparable to the neighborhood graph of the CSV. The network visualization has a multiplicative parameter (c) to adjust its edges visibility. Due to the medoid-based, in addition, it is more an appropriate visualization technique for any type of data than a neighborhood graph of the CSV
A survey of graph-based algorithms for discovering business processes
Full text linkAlgorithms of process discovery help analysts to understand business processes and problems in a system by creating a process model based on a log of the system. There are existing algorithms of process discovery, namely graph-based. Of all algorithms, there are algorithms that process graph-database to depict a process model. Those algorithms claimed that those have less time complexity because of the graph-database ability to store relationships. This research analyses graph-based algorithms by measuring the time complexity and performance metrics and comparing them with a widely used algorithm, i.e. Alpha Miner and its expansion. Other than that, this research also gives outline explanations about graph-based algorithms and their focus issues. Based on the evaluations, the graph-based algorithm has high performance and less time complexity than Alpha Miner algorithm
The use of radial basis function and non-linear autoregressive exogenous neural networks to forecast multi-step ahead of time flood water level
Full text linkMany different Artificial Neural Networks (ANN) models of flood have been developed for forecast updating. However, the model performance, and error prediction in which forecast outputs are adjusted directly based on models calibrated to the time series of differences between observed and forecast values, are very interesting and challenging task. This paper presents an improved lead time flood forecasting using Non-linear Auto Regressive Exogenous Neural Network (NARXNN), which shows better performance in term of forecast precision and produces minimum error compared to neural network method using Radial Basis Function (RBF) in examined 12-hour ahead of time. First, RBF forecasting model was employed to predict the flood water level of Kelantan River at Kuala Krai, Kelantan, Malaysia. The model is tested for 1-hour and 7-hour ahead of time water level at flood location. The same analysis has also been taken by NARXNN method. Then, a non-linear neural network model with exogenous input promoted with enhancing a forecast lead time to 12-hour. Both about the performance comparison has briefly been analyzed. The result verified the precision of error prediction of the presented flood forecasting model
Modified balanced random forest for improving imbalanced data prediction
Full text linkThis paper proposes a Modified Balanced Random Forest (MBRF) algorithm as a classification technique to address imbalanced data. The MBRF process changes the process in a Balanced Random Forest by applying an under-sampling strategy based on clustering techniques for each data bootstrap decision tree in the Random Forest algorithm. To find the optimal performance of our proposed method compared with four clustering techniques, like: K-MEANS, Spectral Clustering, Agglomerative Clustering, and Ward Hierarchical Clustering. The experimental result show the Ward Hierarchical Clustering Technique achieved optimal performance, also the proposed MBRF method yielded better performance compared to the Balanced Random Forest (BRF) and Random Forest (RF) algorithms, with a sensitivity value or true positive rate (TPR) of 93.42%, a specificity or true negative rate (TNR) of 93.60%, and the best AUC accuracy value of 93.51%. Moreover, MBRF also reduced process running time
A comparison on classical-hybrid conjugate gradient method under exact line search
Full text linkOne of the popular approaches in modifying the Conjugate Gradient (CG) Method is hybridization. In this paper, a new hybrid CG is introduced and its performance is compared to the classical CG method which are Rivaie-Mustafa-Ismail-Leong (RMIL) and Syarafina-Mustafa-Rivaie (SMR) methods. The proposed hybrid CG is evaluated as a convex combination of RMIL and SMR method. Their performance are analyzed under the exact line search. The comparison performance showed that the hybrid CG is promising and has outperformed the classical CG of RMIL and SMR in terms of the number of iterations and central processing unit per time
Hierarchical multi-label news article classification with distributed semantic model based features
Full text linkAutomatic news categorization is essential to automatically handle the classification of multi-label news articles in online portal. This research employs some potential methods to improve performance of hierarchical multi-label classifier for Indonesian news article. First potential method is using Convolutional Neural Network (CNN) to build the top level classifier. The second method could improve the classification performance by calculating the average of the word vectors obtained from distributed semantic model. The third method combines lexical and semantic method to extract documents features, which multiplied word term frequency (lexical) with word vector average (semantic). Model build using Calibrated Label Ranking as multi-label classification method, and trained using Naïve Bayes algorithm has the best F1-measure of 0.7531. Multiplication of word term frequency and the average of word vectors were also used to build this classifiers. This configuration improved multi-label classification performance by 4.25%, compared to the baseline. The distributed semantic model that gave best performance in this experiment obtained from 300-dimension word2vec of Wikipedia’s articles. The multi-label classification model performance is also influenced by news’ released date. The difference period between training and testing data would also decrease models’ performance