Search

A dragline is the largest mobile equipment on earth, and it is called the “kingpin” of any mine site. In this present investigation, a case of premature failure of a coal-handling dragline is discussed. Failure occurred in a linked component known as “thimble,” which connects the A-frame and the vertical mast of the dragline. The thimble was broken into two halves and caused the entire front structures of the dragline to collapse. Failure investigation was performed through visual examination, chemical analysis, metallography, mechanical property evaluation and fractography.

one of the greatest health problems in the world is breast cancer. If these breast cancer abnormalities are identified early, there is a maximum chance of recovery. We can go for this early prediction. It is one of the most effective detection and screening strategies and is widely used. The basic goal of CAD systems is to support physicians in the process of diagnosis. CAD systems, however, are very expensive. Our emphasis is on developing a CAD system that is low-cost and effective. To categorize breast cancer as either benign or malignant, a computer-aided detection approach is suggested. The standard mammogram image corpus, Digital Database used for Screening Mammography, images are used for enhancement, segmented and GLCM, intensity and histogram methods are used to extract features. The work is carried out by effective multilayer perceptron classifier (MLP) and support vector machine (SVM). Compare the performance of the classifiers. The proposed approach achieved 96 % accuracy and 8% improvement in accuracy compared to previous approaches with same dataset [4].

Recent studies show that heart attack is one of the severe problems in today’s world. Prediction is one of the crucial challenges in the medical field. In the heart, there are two main blood vessels for the supply of blood through coronary arteries. If the arteries get completely blocked, then it leads to a heart attack. The healthcare field has lots of data related to different diseases, so machine learning techniques are useful to find results effectively for predicting heart diseases. In this paper, data was preprocessed in order to remove the noisy data, filling the missing values using measures of central tendencies. Later, the refined dataset was classified using classifiers apart from prediction. The numbers of attributes were reduced using dimensionality reduction techniques namely Linear Transformation Techniques (LTT) like Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA). The performances of the classifiers were analyzed based on various accuracy-related metrics. The designed classifier model is able to predict the occurrence of a heart attack. The Support Vector Machine (SVM) classifier was applied along with the three kernels namely Linear (linear), Radial Basis Function (RBF), and Polynomial (poly). Another technique namely Decision Tree (DT) was also applied on the Cleveland dataset, and the results were compared in detail and effective conclusions were drawn from the results.

Image denoising is an important preprocessing step done with MRI images to remove various kinds of noise like speckle noise, Gaussian noise, pepper and salt noise, etc. Some filtering mechanisms have been eliminating the required parts of the image along with the noisy pixels of the image, a phenomenon called over-filtering. Anisotropic diffusion is a denoising technique having an iterative process that computes a set of functions to acquire a good degree of smoothening without loss of actual contents of the images. A filtering technique using anisotropic diffusion and application of fuzzy logic has been presented in this paper as it has given a better sharpness of the image, with a good PSNR while it was simulated over 33 MRI cardiac images.

The medium for recording data and information in the present world is paper, magnetic tapes, hard disks, pen drives, etc., whereas about 700 years ago palm leaves were used for this purpose. To recognize the palm leaf text, a novel concept of using a 3D inherent feature, i.e., (depth of incision) is proposed in the current study. This proposed depth sensing approach is used for background subtraction on palm leaf manuscripts. For various features extracted from the palm leaf characters, an improved recognition accuracy is also reported with the help of this 3D feature. To improve the predictive recognition accuracy and to reduce the memory needed, investigations are carried out by implementing optimization techniques.

Ancient wisdom and heritage of Southeast Asian countries were preserved in thousands of palm leaf manuscripts. Due to various factors like aging, insect bites, stains, etc., they are easily susceptible to deterioration. Hence preserving and digitizing such fragile documents is highly essential. Traditional scanning or camera-capturing of such documents suffer from multiple noise artifacts. A depth sensing approach is proposed to eliminate background noise for such manuscripts. The segmented characters extracted from Telugu palm scripts are further recognized using statistical approaches. The improved recognition accuracy is reported using the 3D feature (depth).

Heart disease is the most prevalent kind of disease even among the young people, which has logged the maximum number of lives so far. Earlier it was a disease which claimed people above 50 years of age, but as of now it has it is consuming even the lives of small children with congenital heart diseases. This part of the work also has focused on the cardiac MRI images related to congenital disorders among children in the age group of 2 years to 17 years of age. Medical image enhancement technique is the need of the hour for quick and accurate diagnosis of heart ailments and for medical intervention to take over accordingly. In this paper we focused more on the morphological operations in continuation of our research work in our earlier paper. Different morphological operations were tried over the pre-processed images and the experimental results showed that the opening operation gave a least error measure of all. Various error estimates over these operations were taken, like MSE, RMSE, MAE, etc. were calculated for all MR images and was found the opening operation had a less error compared to other operations, which implies that opened images have more clarity than the other operations.

Characterizing program behavior using static analysis is a challenging problem. In this work, we focus on the fundamental problem of program similarity quantification, i.e., estimating the behavioral similarity of two programs. The solution to this problem is a sub-routine for many important practical problems, such as malware classification, code-cloning detection, program testing, and so on. The main difficulty is to be able to characterize the run-time program behavior without actually executing the program or performing emulation. In this work, we propose a novel behavior tracing approach to characterize program behaviors. We use the call-dependency relationship among the program API calls to generate a trace of the API calling sequence. The dependency tracking is done in a backward fashion, so as to capture the cause and effect relationship among the API calls. Our hypothesis is that this relationship can capture the program behavior to a large extent. We performed experiments by considering several “versions” of a given software, where each version was generated using the code obfuscation techniques. Our approach was found to be resilient up to 20 % obfuscation, i.e., our approach correctly detected that all obfuscated programs that are similar in behavior based on the API call sequences.

Data mining techniques are used for mining useful trends or patterns from textual and image data sets. Medical data mining is very important field as it has significant utility in healthcare domain in the real world. The mining techniques can help the healthcare industry to improve quality of services and grow faster with state-of-the-art technologies. Technology usage is not limited to decision making in enterprises, but spread to every walk of life in all fields. This paper is focused on brain tumour detection which is an essential decision making feature and is a part of healthcare application. This paper proposed a methodology for brain tumour detection which has both training and testing phases. A prototype application has been built to demonstrate the usefulness of the proposed algorithm. The experimental results reveal that the application can be integrated with decision support systems in healthcare domain for improving quality of services.

The Spatial Data is growing very fast but the available statistical techniques are not sufficient to analyze. The existing Spatial Data Mining Techniques also has certain limitations. The size and complexity of the data sets are posing challenges to the research community. In order to overcome these it is required to do deep study on the suitability of the existing Machine Learning Techniques apart from that check for the suitability of hybrid machine learning techniques. In our paper Classifier Ensembling Technique called AdaBoost Approach was applied on the Spatial Data set for rigorous Analysis. The AdaBoost Technique combines multiple weak classifiers into a single Strong Classifier. It is used in conjunction with many machine learning classifier algorithms in order to boost up their performances. In this connection various Tree Classifier Techniques like J48, Random Forest, BF Tree, F Tree, REP Tree, Random Tree, Simple Cart etc., were considered and applied on the Spatial Data set considered and did the comparative study in terms of various performance metric values both in terms of Numerically and Visually and finally made effective conclusions out of that study. This paper also states that ensemble methods perform in better way than any individual classifier.

Background/Objectives: The cases related to Brain Tumor has increased with respect to time owing to various reasons. One of the major challenging issues can be defined by integrating image processing along with data mining algorithms such as classification and clustering. Methods/Statistical Analysis: Artificial Intelligence and Machine Learning techniques are very useful for identifying and visualizing the tumor in the MRI brain image, which can be classified using Support Vector Machines (SVM). Findings: In this paper, we proposed SVM algorithm for classifying the images into two categories Benign and Malignant. SVM classifier model is implemented with RBF and SVM Kernels like linear and non-linear techniques. Application/Improvements: We also proposed for enhancing SVM based MRI Brain image classification by identifying the location and size of the tumor by using different segmentation techniques. The experiments are conducted for evaluating accuracy and the results are compared with existing and proposed work.

Objectives: Ages back mankind depends on agriculture yield and it is the only source for food, income and wealth. Even today people of countries like India depend majorly on Agriculture and allied sectors for their livelihood. Most of India’s income source is from the agriculture sector. Agriculture yield estimation and analysis are not taking place effectively. Method/Analysis: In this regard an algorithm to train the SVM’s i.e., the Sequential Minimal Optimization (SMO), classifier algorithm was proposed and results showed that classifier accuracies were improved when compared to other existing techniques. The process involves in replacing all missing values globally. This implementation is globally and then changes nominal attributes to binary form. By default all the attributes are normalized. Findings: Classifier coefficients output is purely from normalized data rather than from original data, which is very useful and important. Pair wise coupling is a multi-class classification method. Approach addresses the predicted probabilities that are coupled with the pairwise coupling method of Hastie and Tibshirani’s. The accuracies were very low when 10 fold cross validation is applied. Novelty/Improvement: The pre-processing techniques were enhanced to further improve the performance accuracies of SMO algorithm even when cross fold validation is applied on the data sets. Performance based com-parisons were made with the existing techniques.

Background: The availability of Spatial Data which is a part of GIS is growing day by day in an exponential manner. This high availability of data is throwing challenges to the research community to analyze and draw effective conclusions. The Present Study aims at requirement for effective analysis and to draw Conclusions. Methods/Statistical Analysis: Spatial Analysis requires logical relationships between attribute data and map features.Spatial data Analysis is not a simple single task it requires complex procedures in which combinational techniques namely Hybrid techniques are required for effective analysis. Mathematics and statistics are the fundamentals to spatial data analytics. In this paper, a realistic Spatial crime data set was considered for analysis. It involves different types of data mining Techniques like Clustering, Classification and Association rule mining techniques apart from Hybrid techniques. These hybrid Data mining Techniques were applied using R. Findings: The Hybrid Data Mining techniques with K-means Clustering and J48 Decision Tree Algorithm was developed and Applied for the enhancement of accuracy. Association Rule generation Apriori algorithm was applied on the resultant K-means clustered data set. The application of 3D visualization techniques also made for further analysis. Applications/Improvements: It is essentially required to analyze these complex spatial data sets effectively. So there is a need of hybrid Spatial Data Mining Techniques requirement for effective analysis and to draw Conclusions.

In this paper we introduced the three image processing filtering methods. It will be useful for the detection of micro calcification in mammogram image analysis. The preprocessing method plays a vital role in automatic detection of breast cancer. The introduced three filtering methods are Median, Bilateral and the Inverse transform filter. The filter performance analyzed based on the results produced with the help of PSNR and MSE values. Experimental results found that the Inverse transform filter has given PSNR= 95.37, MSE= 0.00001 with noise density 0.005 for salt-and-pepper, Gaussian, Poisson and speckle noise. Then we can say that this filter is the one of the best filters for medical applications.

The World Wide Web (WWW) allows the people to share the information (data) from the large database repositories globally. The amount of information grows billions of databases. We need to search the information will specialize tools known generically search engine. There are many of search engines available today, retrieving meaningful information is difficult. However to overcome this problem in search engines to retrieve meaningful information intelligently, semantic web technologies are playing a major role. In this paper we present survey on the search engine generations and the role of search engines in intelligent web and semantic search technologies .

This paper deals with the recognition of Telugu characters on palm leaf using statistical features. Handwritten character recognition has various applications in post offices, reading aids for blind, library automation and multimedia design. Palm leaf manuscripts contain religious texts and a host of subjects such as art, medicine, music, astrology, law and astronomy. There is an inherent 3D feature for characters on palm leaf called depth. This depth is proportional to the writers stylus pressure applied at each pixel point. This 3D feature of every pixel in an image is used to recognize the palm leaf characters in the present work. The image is divided into zones and the sum of the pixel intensities in each zone is used as a feature vector to recognize the palm leaf characters. As per the literature survey, the recognition accuracy for handwritten characters is less than 60% and also very less amount of work is done for palm leaf character recognition. Using the proposed method the best recognition accuracy obtained for palm leaf characters is 96%.

This paper focuses on early stage lung cancer detection. Lung cancer is prominent cancer as it states large number of deaths of more than a million every year. It creates need of detecting the lung nodule at early stage in Computer Tomography medical images. So to detect the occurrence of cancer nodule at early stage, the requirement of methods and techniques is increasing. There are different methods and techniques existing but none of them provide a better accuracy of detection. One of the techniques is content based image retrieval Computer Aided Diagnosis System (CAD) for early detection of lung nodules from the Chest Computer Tomography (CT) images. This optimization algorithm allows physicians to identify the nodules present in the CT lung images in the early stage hence the lung cancer. The MATLAB image processing toolbox based implementation is done on the CT lung images and the classifications of these images are carried out. The performance measures like the classification rate and the false positive rates are analyzed.

Character recognition is one of the oldest applications of pattern recognition. Recognizing Hand-Written Characters (HWC) is an effortless task for humans, but for a computer it is a difficult job. Research in character recognition is very popular for various potential applications such as in banks, post offices, defense organizations, reading aid for the blind, library automation, language processing and multi-media design. Optical Character Recognition (OCR) is based on optical mechanism which consists of a machine to recognize scanned and digitized character automatically. Automatic recognition of handwritten text can be done either Offline or Online. Offline handwritten recognition is the task of recognizing the image of a hand written text, in contrast to Online recognition where the dynamic characteristics of the writing are available and recorded while the scriber is writing on a special screen with a pen/stylus made for this application. Zonal based feature extraction is used in the present proposed method. The character image is divided into predefined number of zones and a statistical feature is computed from each of these zones. Usually, this feature is based on the pixels contained in that zone. The gray values of the pixels in that selected zone are summed up to form a feature for that zone in that image. The features of all the zones in the image form a feature vector which is used for handwritten character recognition. In this work, using this Zoning method the recognition accuracy is found to be 78%.

The main idea of this research is to first design the similarity measure which can be used to of find the similarity between any two text documents and use the same to perform clustering. The similarity measure designed is analyzed to study the behavior in the best case, average case and worst case situations. The drawback of Euclidean, Cosine, Jaccard similarity measures are overcome using the proposed measure. The similarity measure is evaluated considering reuters-21578 dataset. The results show that the proposed measure overcomes other measures.

Study of remote sensed imagery has gained practical significance in various domains such as environmental monitoring, fire risk mapping, change detections and land use. Classification is a data mining methodology which is used to assign class labels to data instances and build a model so as to be able to predict class labels for unlabelled data. In this paper algorithms based on parametric distribution model like k nearest neighbor classifier and linear kernel based support vector machines classifier are used for classifying remote sensed data. A generic algorithm is discussed to implement the said classification. We finally analyze the performance of these algorithms based on various parameters.

Environmental monitoring is one of the key approaches to safeguard the global ecosystem. Classifications of different water levels facilitate in preserving water reserves and maintain the equilibrium in the ecosystem. In this paper we shall inspect the classification of drainage water levels in Canada. A powerful statistical tool called support vector machines is used to classify the said drainage water remote sensed spatial data sets. To boost the performance of support vector machines classifier a new generic algorithm based on parametric distribution model will be proposed. Later several evaluation metrics like kappa statistics are used to compare the results of the proposed algorithm with multi-layer perceptron neural networks and naive bayes classifiers.

The problem of missing data in the real world datasets has very significant role in the real time data mining process and becomes more complex in large databases. The presence of missing values influences data set features and the class attributes, thus affecting the predictive accuracies of the classifiers. For the last one decade, many researchers have come out with different techniques for dealing with missing attribute values in databases with homogeneous and/or numeric attributes. In this research work, we proposed a new indexing measure to the imputation algorithm for missing data values of the attributes to compute the similarity measure between any two typical elements in the dataset. It can also be applied on any dataset be it a nominal and/or real. The proposed algorithm is evaluated by extensive experiments and comparison with KNNI, SVMI, WKNNI, KMI and FKMI algorithms. The results showed that the proposed algorithm has better performance than the existing imputation algorithms in terms of classification accuracy and also our decision tree algorithm employs highly accurate decision rules.

The aim of this research paper is the classification of magnetic resonance brain Images as normal and abnormal using linear kernel based Support Vector Machine (SVM) for which different texture features are utilized to characterize the information level contained in the image. The proposed method is compared with k-nearest neighbor (K-NN) classifier and hidden markov model (HMM) classifier. To evaluate the performance of the proposed method, classification rate, recall, and precision evaluation metrics are choosen. The comparative results of the research demonstrates that SVM based on linear kernel provides much higher precision and low error rates as compared to KNN and HMM classifier.

Research in optical character recognition (OCR) had started more than five decades ago. The recognition accuracy for printed characters is above 90%, whereas for handwritten characters is very low and less than 60% as reported in the literature. Handwritten character recognition of Indian languages is still at nascent stage of research and hence captivated our attention for further analysis. This paper describes the handwritten character recognition of Telugu language using two-stage classifiers. k-Nearest Neighbor (k-NN) and Support Vector Machines (SVM) were used for classification in this work. The use of these two classifiers one after the other in two subsequent stages increases the recognition accuracy of the system. Various features extracted from the images are block pixel count, block based directions, histograms and boundaries for both the training and test images. In the first stage k-NN classifier was used and subsequently the wrongly recognized characters were tested with SVM classifier. Again the classifiers were interchanged in the first and second stages to check the improvement of accuracy. It was found that the recognition accuracy increased to a great extent by cascading the two different classifiers. Using these two classifiers the best recognition accuracy obtained was 90.2%.