Informatica

Estimation of the ratio of two totals is considered, when a probability sample from the finite population is available. Four estimators of the ratio are examined. The first one – called “simple” – is the ratio of the Horvitz–Thompson estimators of totals; the second – the ratio of two ratio estimators of totals; the third one – the ratio of two regression estimators of totals. The fourth one is a calibrated estimator of the ratio. The variances of these estimators are compared. The properties of such estimators of the ratio are studied. The simulation results are presented.

In this paper, a reliable information hiding scheme based on support vector machine and error correcting codes is proposed. To extract the hidden information bits from a possibly tampered watermarked image with a lower error probability, information hiding is modeled as a digital communication problem, and both the good generalization ability of support vector machine and the error correction code BCH are applied. Due to the good learning ability of support vector machine, it can learn the relationship between the hidden information and corresponding watermarked image; when the watermarked image is attacked by some intentional or unintentional attacks, the trained support vector machine can recover the right hidden information bits. The reliability of the proposed scheme has been tested under different attacks. The experimental results show that the embedded information bits are perceptually transparent and can successfully resist common image processing, jitter attack, and geometrical distortions. When the host image is heavily distorted, the hidden information can also be extracted recognizably, while most of existing methods are defeated. We expect this approach provide an alternative way for reliable information hiding by applying machine learning technologies.

The conjugate gradient method is an iterative technique used to solve systems of linear equations. The paper analyzes the performance of parallel preconditioned conjugate gradient algorithms. First, a theoretical model is proposed for estimation of the complexity of PPCG method and a scalability analysis is done for three different data decomposition cases. Computational experiments are done on IBM SP4 computer and some results are presented. It is shown that theoretical predictions agree well with computational results.

This paper describes our research on statistical language modeling of Lithuanian. The idea of improving sparse n‐gram models of highly inflected Lithuanian language by interpolating them with complex n‐gram models based on word clustering and morphological word decomposition was investigated. Words, word base forms and part‐of‐speech tags were clustered into 50 to 5000 automatically generated classes. Multiple 3‐gram and 4‐gram class‐based language models were built and evaluated on Lithuanian text corpus, which contained 85 million words. Class‐based models linearly interpolated with the 3‐gram model led up to a 13% reduction in the perplexity compared with the baseline 3‐gram model. Morphological models decreased out‐of‐vocabulary word rate from 1.5% to 1.02%.

Text categorization – the assignment of natural language documents to one or more predefined categories based on their semantic content – is an important component in many information organization and management tasks. Performance of neural networks learning is known to be sensitive to the initial weights and architecture. This paper discusses the use multilayer neural network initialization with decision tree classifier for improving text categorization accuracy. Decision tree from root node until a final leave is used for initialization of each single unit. Growing decision trees with increasingly larger amounts of training data will result in larger decision tree sizes. As a result, the neural networks constructed from these decision trees are often larger and more complex than necessary. Appropriate choice of certainty factor is able to produce trees that are essentially constant in size in the face of increasingly larger training sets. Experimental results support the conclusion that error based pruning can be used to produce appropriately sized trees, which are directly mapped to optimal neural network architecture with good accuracy. The experimental evaluation demonstrates this approach provides better classification accuracy with Reuters‐21578 corpus, one of the standard benchmarks for text categorization tasks. We present results comparing the accuracy of this approach with multilayer neural network initialized with traditional random method and decision tree classifiers.

Walras theory is well known and widely used in models of market economy. Various iterative methods are developed to search for the equilibrium conditions.

Tree is one of the most studied and practically useful classes of graphs and is the attention of a great number of studies. There is absence of generalized results for tree as a class and even for one kind of labeling as whole. Only specialized results exist limited to specific types of trees only. A number of conjectures stand being unsolved. Graham and Sloane (1980) conjectured trees to be Harmonious and Ringel‐Kotzig conjectured trees to be Graceful about three decades ago. Kotzig and Rosa (1970) ask the question whether all trees are Magic or not. No generalized result for Antimagic labeling is given for trees so far. This paper presents the methodologies to obtain the major labeling schemes for trees viz., Harmonious, Sequential, Felicitous, Graceful, Antimagic and found the trees to be not Magic except T(2,1), thus solving the said conjectures. These findings could also be useful for those working in fields where graphs serve as models.

The paper considers moving locally predefined (MLP) finite element remeshing technique for deep penetration of the rigid cone into homogeneous and porous medium. Remeshing presents a computational tool implemented in the form of postprocessor type software compatible with standard FEM codes. It involves a transfer operation combining both the moving least square method based on stress patch recovery and the interpolation method for transfer of state variables. The developed MLP remeshing is able to overcome numerical difficulties occurring due to large distortions of the Lagrangian mesh and contact sliding and capture steady‐state behavior. It shows good performance in modeling of cone penetration into elasto‐plastic homogeneous and porous media reaching several diameters of the cone.

In this paper, the main measure, an amount of information, of the information theory is analyzed and corrected. The three conceptions of the theory on the microstate, dissipation pathways, and self‐organization levels with a tight connection to the statistical physics are discussed. The concepts of restricted information were introduced as well as the proof of uniqueness of the entropy function, when the probabilities are rational numbers, is presented.

The development of Lithuanian HMM/ANN speech recognition system, which combines artificial neural networks (ANNs) and hidden Markov models (HMMs), is described in this paper. A hybrid HMM/ANN architecture was applied in the system. In this architecture, a fully connected three‐layer neural network (a multi‐layer perceptron) is trained by conventional stochastic back‐propagation algorithm to estimate the probability of 115 context‐independent phonetic categories and during recognition it is used as a state output probability estimator. The hybrid HMM/ANN speech recognition system based on Mel Frequency Cepstral Coefficients (MFCC) was developed using CSLU Toolkit. The system was tested on the VDU isolated‐word Lithuanian speech corpus and evaluated on a speaker‐independent ∼750 distinct isolated‐word recognition task. The word recognition accuracy obtained was about 86.7%.