The basic solution for locating an optimal reduct is to generate all possible reducts and select the one that best meets the given criterion. Since this problem is NP-hard, most attribute reduction algorithms use heuristics to find a single reduct with the risk to overlook for the best ones. There is a discernibility function (DF)-based approach that generates all reducts but may fail due to memory overflows even for datasets with dimensionality much below the medium. In this study, we show that the main shortcoming of this approach is its excessively high space complexity. To overcome this, we first represent a DF of \(n\) attributes by a bit-matrix (BM). Second, we partition the BM into no more than \(n-1\) sub-BMs (SBMs). Third, we convert each SBM into a subset of reducts by preventing the generation of redundant products, and finally, we unite the subsets into a complete set of reducts. Among the SBMs of a BM, the most complex one is the first SBM with a space complexity not greater than the square root of that of the original BM. The proposed algorithm converts such a SBM with \(n\) attributes into the subset of reducts with the worst case space complexity of \(\left( _{n/2}^n \right) /2\).

Protein-protein interactions hold very important roles in biological processes. Prediction of PPIs is important for understanding these processes. In this context, substantive representations of proteins are needed during the process of interaction prediction in order to achieve higher prediction accuracy. In this paper, a new feature representation method, based on the concept of Chou's pseudo amino acid composition, was introduced. It is composed of the weighted amino acid composition information and the correlation factors of the protein. Finally, an SVM classification model was constructed for predicting PPIs. Experimental results exhibit that our method precedes those previously published in the literature.

Logic minimization is used for solving many problems such as design of logic circuits in logic synthesis and attribute reduction, rule induction and query optimization in information systems. Since exact logic minimization is a NPhard problem, usually logic minimization problems are solved by the heuristics approaches among which the most frequently used one is the sequential covering strategy. In our opinion, the most efficient approach to realization of this strategy is the reduced off-set based-direct cover minimization developed for logic synthesis. But this approach works in the search space of size 4 that makes the minimization process very complex. We show that this problem may be reduced in a large scale by reducing the search space from 4 to 2. But this causes to generating the results in an incompletely specified form. We show that this problem may be solved by a simple rule.

Usually a dataset has a lot of reducts finding all of which is known to be an NP hard problem. On the other hand, different reducts of a dataset may provide different classification accuracies. Usually, for every dataset, there is only a reduct with the best classification accuracy to obtain this best one, firstly we obtain the group of attributes that are dominant for the given dataset by using the decision tree algorithm. Secondly we complete this group up to reducts by using discernibility function techniques. Finally, we select only one reduct with the best classification accuracy by using data mining classification algorithms. The experimental results for datasets indicate that the classification accuracy is improved by removing the irrelevant features and using the simplified attribute set which is derived from proposed method.

Anahtarlama fonksiyonlarinin sadelestirilmesi tasarimcilara daha kisa zaman suresinde, daha sade lojik devrelertasarlama imkani saglamaktadir. Sadelestirilmis olan bir fonksiyon daha az guc tuketimi, daha az hacim ve dahaaz maliyet gerektirir. Bu konu ile ilgili olarak gelistirilen yontemlerin cogu iki ana adimda gerceklestirilir.Birinci adimda, asal carpim terimlerinin tumu belirlenir. Ikinci adimda fonksiyonu sadelesmis olarak ortecek,esas asal carpim terimler kumesi belirlenir. Anahtarlama fonksiyonlarini sadelestirecek algoritmalarin tumuO(2n) karmasikligina sahiptirler. Arastirmalar gostermistir ki n’in cok yuksek degerlerinde esas asal carpimterimlerin tam kumesini belirleme yontemi pratik olarak gerceklestirilemez duruma gelmektedir. Bu yuzden bucalismada, asal carpim terimlerin belli kistaslara cevap verecek alt kumeleri olusturularak, dogrudan ortmeprensibine dayanan yakin minimum sadelestirme algoritmasi gelistirilmistir. Gelistirilen algoritma cesitli problemleruzerinde test edilmis ve dunyaca ornek olarak kabul edilen ESPRESSO algoritmasi ile karsilastirilmistir.Karsilastirma kistaslari olarak algoritmalarin; cozum sonucunda bulduklari carpim terimlerinin toplamifadelerinin sayisi, cozume ulasma sureleri ve cozume ulasirken kullandiklari bellek kapasitesi alinmistir.Karsilastirma sonuclarina gore gelistirilen algoritmanin basarili sonuclar verdigi gorulmustur.

In order to generate prime implicants for a given cube (minterm), most of minimization methods increase the dimension of this cube by removing one literal from it at a time. But there are two problems of exponential complexity. One of them is the selection of the order in which the literals are to be removed from the implicant at hand. The latter is the mechanism that checks whether a tentative literal removal is acceptable. The reduced Offset concept has been developed to avoid of these problems. This concept is based on positional-cube representation where each cube is represented by two n-bit strings. We show that each reduced Off-cube may be represented by a single n-bit string and propose a set of bitwise operations to be performed on such strings. The experiments on single-output benchmarks show that this approach can significantly speed up the minimization process, improve the quality of its results and reduce the amount of memory required for this aim.

Real life data sets often contain noisy data which makes the subsequent data mining process difficult. The feature selection preprocessing step can be simplified the datasets by eliminating the features that are redundant for classification process, with pertinent features would reduce the size of dataset and afterwards allow more apparent analysis of extracted rules pattern and rules. This paper introduces a method called logic function-based feature selection (L.FB.F.S) for improving the accuracy of data mining classification algorithms. The goal of the feature selection is to find the minimal subsets of attributes of the dataset that can be used for classification tasks by removing both the irrelevant and redundant features. Primarily, L.FB.F.S finds the all MSs of a dataset. Secondly, one of MS with the best classification ability is selected for improving the accuracy of data mining classification algorithms. Key-Words: Feature selection, attribute reduction, classification, Boolean function

Obtaining the association between complex diseases and single nucleotide polymorphisms (SNPs) is one of the most important medical problems. Although obtaining the full set of SNPs is a very challenging issue, there are subsets of tag SNPs, each of which allows predicting the rest of SNPs with enough accuracy. Here, the problem is to obtain such a subset of tag SNPs that makes it possible to predict the rest of SNPs with maximal possible accuracy. However, the methods developed for this aim cannot reach the accuracy level enough for practical applications. In this study, a new approach using the Genetic Algorithm for selecting the tag SNPs and Support Vector Machine for predicting the values of the rest of SNPs is proposed. The results of the experiments performed on a number of datasets demonstrate that the proposed method can predict the values of the rest of SNPs with a significantly better accuracy than other methods with the same purpose.

The goal of attribute reduction is to find a minimal subset (MS) R of the condition attribute set C of a dataset such that R has the same classification power as C. It was proved that the number of MSs for a dataset with n attributes may be as large as (n/2n) and the generation of all of them is an NP-hard problem. The main reason for this is the intractable space complexity of the conversion of the discernibility function (DF) of a dataset to the disjunctive normal form (DNF). Our analysis of many DF-to-DNF conversion processes showed that approximately (1-2/(n/2n) \times 100)% of the implicants generated in the DF-to-DNF process are redundant ones. We prevented their generation based on the Boolean inverse distribution law. Due to this property, the proposed method generates 0.5 \times (n/2n) times fewer implicants than other Boolean logic-based attribute reduction methods. Hence, it can process most of the datasets that cannot be processed by other attribute reduction methods.

The goal of feature selection (FS) is to find the minimal subset (MS) R of condition feature set C such that R has the same classification power as C and then reduce the dataset by discarding from it all features not contained in R. Usually one dataset may have a lot of MSs and finding all of them is known as an NP-hard problem. Therefore, when only one MS is required, some heuristic for finding only one or a small number of possible MSs is used. But in this case there is a risk that the best MSs would be overlooked. When the best solution of an FS task is required, the discernibility matrix (DM)-based approach, generating all MSs, is used. There are basically two factors that often cause to overflow the computer’s memory due to which the DM-based FS programs fail. One of them is the largeness of sizes of discernibility functions (DFs) for large data sets; the other is the intractable space complexity of the conversion of a DF to disjunctive normal form (DNF). But usually most of the terms of DF and temporary results generated during DF to DNF conversion process are redundant ones. Therefore, usually the minimized DF (DFmin) and the final DNF is to be much simpler than the original DF and temporary results mentioned, respectively. Based on these facts, we developed a logic function-based feature selection method that derives DFmin from the truth table image of a dataset and converts it to DNF with preventing the occurrences of redundant terms. The proposed method requires no more amount of memory than that is required for constructing DFmin and final DNF separately. Due to this property, it can process most of datasets that can not be processed by DM-based programs.

The goal of attribute reduction is to reduce the problem size and search space for learning algorithms. The basic solution of this problem is to generate all possible minimal attributes subsets (MASes) and choose one of them, with minimal size. This can be done by constructing a kind of discernibility function (DF) from the dataset and converting it to disjunctive normal form (DNF). Since this conversion is NP-hard, for attribute reduction usually heuristic algorithms are used. But these algorithms generate one or a small number of possible MASes that generally is not sufficient for optimality of dataset processing in such aspects as the simplicity of data representation and description, the speed and classification accuracy of the data mining algorithms and the required amount of memory. In this study, we propose an algorithm that finds all MASes by iteratively partitioning the DF so that the part to be converted to DNF in each of iterations has the space complexity no higher than the square root of the worst-case space complexity of the conversion of the whole DF to DNF. The number of iterations is always fewer than the number of attributes.

This study analyzes the effect of discretization on classification of datasets including continuous valued features. Six datasets from UCI which containing continuous valued features are discretized with entropy-based discretization method. The performance improvement between the dataset with original features and the dataset with discretized features is compared with k-nearest neighbors, Naive Bayes, C4.5 and CN2 data mining classification algorithms. As the result the classification accuracies of the six datasets are improved averagely by 1.71% to 12.31%. Keywords—Data mining classification algorithms, entropy-based discretization method

Since the generation all of minimal subsets of attributes (MSAs) of a dataset is NP-hard, usually attribute reduction algorithms (ARAs) use some heuristics to find a small part of MSAs with the risk to overlook the best solutions. There is a discernibility function (DF)-based ARA for generating all MSAs, but often failing to find them even for medium sized datasets due to memory overflows. In this study, it is proposed such a partition of a DF on which the computational complexity of each part cannot be exceeded the square root of the computational complexity of the DF in whole.

Investigations on genetic variants associated with complex diseases are important for enhancements in diagnosis and treatments. SNPs (Single Nucleotide Polymorphisms), which comprise most of the millions of changes in human genome, are promising tools for disease-gene association studies. On the other hand, these studies are limited by cost of genotyping tremendous number of SNPs. Therefore, it is essential to identify a subset of SNPs that represents rest of the SNPs. As subset of SNPs is identified, data set should be searched as well as possible. In this study, a new method called CLONTagger was introduced, where Support Vector Machine (SVM) was used as SNP prediction method, whereas Clonal Selection Algorithm (CLONALG) was used as tag SNP selection method. The suggested method was compared with current tag SNP selection algorithms in literature using different datasets. Experimental results demonstrated that the suggested method could identify tag SNPs with better prediction accuracy than other methods from literature.

In this study, the effects of hardness on temperature increase of ST 37 steel during fatigue loading were investigated. Steel specimens are made of ST 37 steel and subjected to heat treatment to obtain different hardness. The specimens were subjected to reverse bending fatigue loading and the specimens were observed by using a infrared (IR) camera during the test. The obtained thermal images were recorded by FLIR E45 IR camera and then transferred to the image processing program developed by using MATLAB. Thus after image processing, thermal values used to detect the temperature rise of the surface of the steel specimen under fatigue loading were obtained. During the fatigue, the material is subjected to strain energy input which result in plastic or/and elastic deformation. This event results in an increase of temperature within material. The energy conservation requires that the generated heat shows itself as heat transfer by conduction, convection and radiation and internal energy increase. Besides, if the material has undergone plastic deformation, an additional term which accounts this effect should be included within energy conservation equation. In order to observe the effect of plastic deformation upon temperature increase of material, the ability of plastic deformation has been changed through the change of hardness and the thermal variations during fatigue has been investigated.

In this study, the temperature rise of composite plates with a hole during fatigue loading was investigated. Woven glass/epoxy composite plates with eight plies were subjected to bending fatigue loading and materials were observed by using a thermal camera during the test. Previous works showed that a heat generation can form due to internal friction and damage formation. Therefore, a thermographic infrared imaging system was used to detect the temperature rise of composite specimens. During the tests, the thermal images of the specimens have been recorded by a thermal camera and then transferred to the image processing program which has been developed by using MATLAB. By using these thermal images, the spot temperatures of the specimen were obtained by using artificial neural networks. The obtained temperatures show local increase at places where the heat generation localized. These regions considered being the probable damage initiation sites. It is shown in this study that most probable damage initiation zones in the woven glass/epoxy composite material can be detected by using infrared thermography (IRT) approach prior to failure.

In this study, a novel OFF-set based direct-cover Exact Minimization Algorithm (EMA) is proposed for single-output Boolean functions represented in a sum-of-products form. To obtain the complete set of prime implicants covering the given Target Minterm (ON-minterm), the proposed method uses OFF-cubes (OFF-minterms) expanded by this Target Minterm. The amount of temporary results produced by this method does not exceed the size of the OFF-set. In order to achieve the goal of this study, which is to make faster computations, logic operations were used instead of the standard operations. Expansion OFF-cubes, commutative absorption operations and intersection operations are realized by logic operations for fast computation. The proposed minimization method is tested on several classes of benchmarks and then compared with the ESPRESSO algorithm. The results show that the proposed algorithm obtains more accurate and faster results than ESPRESSO does.

Protein-protein interactions are critical in coordinating various cellular processes. They help understanding protein function and drug design. Extracting protein features from amino acid sequences is important in order to study protein-protein interactions. Various feature extraction approaches for proteins have been introduced up to the present. PseAAC is one of the most used protein feature extractor. In this work we purpose a new approach to calculate amino acid composition values. The purpose of our method is to adjust the weights of the composition values during feature extraction process. It means that bigger composition values will contribute more to prediction function than smaller ones. Our experimental results showed that our method outperformed PseAAC.

In this paper chaotic systems and RSA encryption algorithm are combined in order to develop an encryption algorithm which accomplishes the modern standards. E.Lorenz's weather forecast' equations which are used to simulate non-linear systems are utilized to create chaotic map. This equation can be used to generate random numbers. In order to achieve up-to-date standards and use online and offline status, a new encryption technique that combines chaotic systems and RSA encryption algorithm has been developed. The combination of RSA algorithm and chaotic systems makes encryption system.

The minimization of Switching functions allows designers to make use of fewer components, thus reducing the cost of particular system. Simplified as a function requires less power consumption, less volume and less cost. Most of minimization techniques work on a two-step principle, the first step identifies all of the prime implicants and the second step selects the subset of prime implicants that covers the function(s) being minimized. All procedures for Boolean networks into prime and irredundant form have O(2(n)) complexity. Prime Implicants identification step can be computational impractical as n increases. Therfore, in this study, subsets of prime implicants that can prove direct cover principle which based on definite criterions use for mimimization method. The method has been tested on several different kinds of problems and results of which were compared with ESPRESSO Comparison of algorithms as benchmarks; solution as a result they find that their total number of product terms, the solution times and reach a solution when they reach the memory capacity is taken. According to the results of the comparison developed algorithm gives successful results.