Department of Computer Science (PG)
Permanent URI for this communityhttps://dspace.psgrkcw.com/handle/123456789/146
Browse
11 results
Search Results
Item MISSING VALUE AWARE OPTIMAL FEATURE SELECTION METHOD FOR EFFICIENT BIG DATA MINING PROCESS(International Journal of Recent Technology and Engineering (IJRTE), 2019-09) S, Meera; B, Rosiline JeethaBig mining plays a more critical role in the real world environment due to presence of large volume of data with different varieties and type. Handling these data values and predicting the information would be the more difficult task which needs to be concerned more to obtain the useful knowledge. This is achieved in our previous research work by introducing the Enhanced Particle Swarm Optimization with Genetic Algorithm – Modified Artificial Neural Network (EPSOGA -MANN) which can select the optimal features from the big volume of data. However this research work might be reduced in its performance due to presence of missing values in the dataset. And also this method is more complex to perform due to increased computational overhead of ANN algorithm. This is resolved in the proposed research method by introducing the method namely Missing Value concerned Optimal Feature Selection Method (MV-OFSM). In this research method Improved KNN imputation algorithm is introduced to handle the missing values. And then Dynamic clustering method is introduced to cluster the dataset based on closeness measure. Then Anarchies Society Optimization (ASO) based feature selection approach is applied for performing feature selection in the given dataset. Finally a Hybrid ANN-GA classification technique is applied for implementing the classification. The overall performance evaluation of the research method is performed in the matlab simulation environment from which it is proved that the proposed research method leads to provide the better performance than the existing research techniqueItem A SURVEY OF PARALLEL SOCIAL SPIDER OPTIMIZATION ALGORITHM BASED ON SWARM INTELLIGENCE FOR HIGH DIMENSIONAL DATASETS(International Journal of Computational Intelligence Research, 2017-11-09) B, Shanmugapriya; S, MeeraBig data is the slightly abstract phase which describes the relationship between the data size and data processing speed in the system. The many new information technologies the big data deliver dramatic cost reduction, substantial improvements in the required time to perform the computing task or new product and service offerings. The several complicated specific and engineering problems can be transformed in to optimization problems. Swarm intelligence is a new subfield of computational intelligence (CI) which studies the collective intelligence in a group of simple intelligence. In the swarm intelligence, useful information can be obtained from the competition and cooperation of individuals. In this paper discussed about some of the optimization algorithms based on swarm intelligence such as Ant Colony optimization (ACO), Particle Swarm Algorithm (PSO), Social Spider Optimization (SSO) Algorithm and Parallel Social Spider Optimization (P-SSO) Algorithm. These optimization techniques are based on their merits, demerits and metrics accuracy, sum of intra cluster distance, Recovery Error Etc.Item SURVEY ON SWARM SEARCH FEATURE SELECTION FOR BIG DATA STREAM MINING.(International Journal of Computational Intelligence Research, 2017-01) S, Meera; B, Rosiline JeethaBig data is the slightly abstract phase which describes the relationship between the data size and data processing speed in the system. The many new information technologies the big data deliver dramatic cost reduction, substantial improvements in the required time to perform the computing task or new product and service offerings. The several complicated specific and engineering problems can be transformed in to optimization problems. Swarm intelligence is a new subfield of computational intelligence (CI) which studies the collective intelligence in a group of simple intelligence. In the swarm intelligence, useful information can be obtained from the competition and cooperation of individuals. In this paper discussed about some of the optimization algorithms based on swarm intelligence such as Ant Colony optimization (ACO), Particle Swarm Algorithm (PSO), Social Spider Optimization (SSO) Algorithm and Parallel Social Spider Optimization (P-SSO) Algorithm. These optimization techniques are based on their merits, demerits and metrics accuracy, sum of intra cluster distance, Recovery Error Etc.Item AN ANN APPROACH FOR CLASSIFICATION OF BRAIN TUMOR USING IMAGE PROCESSING TECHNIQUES IN MRI IMAGES(International Journal Of Engineering Sciences & Research Technology, 2015-12) A, Sindhu; S, MeeraA brain tumor is defined as the growth of abnormal cells in the tissues of the brain. Brain tumors can be benign (noncancerous) or malignant (cancerous). MRI represents an interesting approach for the anatomical assessment of brain tumors since it provides superior soft tissue contrast and high-resolution information. MRI scan images are taken for this project to process further. This work proposed artificial neural network approach namely Back propagation network (BP-ANN).Image segmentation is done by using region growing algorithm which is used to detect the tumor present in the brain MRI images. GLCM is used for extracting the brain features in the images. This system proposed two modes namely training and testing phase which is used to classify the output.Item A SURVEY ON DETECTING BRAIN TUMOR IN MRI IMAGES USING IMAGE PROCESSING TECHNIQUES(International Journal of Innovative Research in Computer and Communication Engineering, 2015-01) A, Sindhu; S, MeeraMedical Image Processing is the fast growing and challenging field now a days. Medical Image techniques are used for Medical diagnosis. Brain tumor is a serious life threatening disease. Detecting Brain tumor using Image Processing techniques involves four stages namely Image Pre-Processing, Image segmentation, Feature Extraction, and Classification. Image processing and neural network techniques are used to improve the performance of detecting and classifying brain tumor in MRI images. In this survey various Image processing techniques are reviewed particularly for Brain tumor detection in magnetic resonance imaging. More than twenty five research papers of image processing techniques are clearly reviewed.Item ACCELERATION ARTIFICIAL BEE COLONY OPTIMIZATION-IMPROVED TRANSDUCTIVE SUPPORT VECTOR MACHINE FOR EFFICIENT FEATURE SELECTION IN BIG DATA STREAM MINING.(Jour of Adv Research in Dynamical & Control Systems, 2017-04) S, Meera; B, Rosiline JeethaHigh dimensional data seen in a practical issue imposes a hurdle for large data analysis. Attribute reduction or feature selection aids the learning algorithm to work with efficiency by eliminating unnecessary and repetitive information in the big data. The existing system like Acceleration Particle Swarm Optimization–Support Vector Machine (APSO-SVM) is proposed in order to deal with the above challenge. But the already existing technique has issues in addition to the pre-processing technique and optimal feature selection for scalable dataset. Therefore the system’s overall performance is decreased significantly. With the aim of eliminating these problems, in the proposed system, Acceleration Artificial Bee Colony –Improved Transductive SVM (AABC-ITSVM) is introduced so as to improve the system performance in a more efficient manner. The proposed system comprises of three important modules like preprocessing, feature selection and classification. The preprocessing is carried out by making use of min-max normalization algorithm that assists in increasing the classification accuracy more. Thereafter the feature selection is carried out by employing AABC optimization algorithm that is utilized for selecting the significant and necessary features from the data that is preprocessed. The selected features are classified by employing ITSVM algorithm. The ITSVMs gets the labeling of the test features, which increases the margin conjoined on the training and the test data. It yields classification results with more accuracy for the datasets specified. The proposed system offers great performance with regard to superior accuracy, recall, sensitivity, specificity, precision, f-measure, gmean, and lesser selected features, time complexity by utilizing the AABCITSVM technique.Item ENHANCED PARTICLE SWARM OPTIMIZATION WITH GENETIC ALGORITHM AND MODIFIED ARTIFICIAL NEURAL NETWORK FOR EFFICIENT FEATURE SELECTION IN BIG DATA STREAM MINING(PSG College of Technology, 2019-01-05) S, Meera; B, Rosiline JeethaIn the recent years the volume of data has been growing tremendously with the development in the information technology. Hugh dimensionality would be one of the major challenges faced by people working in research with big data as a high dimensionality that happens while a dataset comprises of a big number of features (autonomous attributes). For resolving this issue, often researchers make use of a feature selection step for identification and removal of irrelevant features (not helpful in the classification of the data) and repetitive features (yield the same information like other features). Acceleration Artificial Bee Colony- Artificial Neural Network (AABC-ANN) has been introduced in the preceding research for handling the feature selection process over the big data. Computational complexity and inaccuracy of dataset remains as a problem for these methods. Enhanced Particle Swarm Optimization with Genetic Algorithm – Modified Artificial Neural Network (EPSOGA -MANN) is proposed in the proposed methodology for avoiding the above mentioned issues. Modules’ including preprocessing, feature selection and classification has been included in this research process. Fuzzy C Means (FCM) denotes the clustering algorithm which is used to handle the noise information efficiently in preprocessing. Issues like missing data, repetitiveness and error data are resolved effectively. Discarding of irrelevant features leads to the significant decrease in size of the structured and semi structured dataset. These features are then considered for the feature selection process. Feature selection process is carried out by means of EPSOGA algorithm optimally in this research. More important and relevant features are selected by EPSOGA optimization algorithm from the as it provides classification results with more accuracy for the huge volume of dataset given. Input, hidden and output layer are the three layers of MANN. It is introduced for improving the time complexity by means of neurons. This proposed methodology is more suitable for handling the big data principles such as volume, velocity and variety. The proposed model provides superior performance as show in experimental results in terms of superior accuracy, recall, precision, f-measure, and lesser time complexity by means of EPSOGA –MANN approachItem ACCELERATION ARTIFICIAL BEE COLONY OPTIMIZATION-ARTIFICIAL NEURAL NETWORK FOR OPTIMAL FEATURE SELECTION OVER BIG DATA(Saveetha Engineering College, Chennai, 2012-09) S, Meera; B, Rosiline JeethaFinding an appropriate set of features from data of high dimensionality for building an accurate classification model is challenging in recent years. In the preceding research, Acceleration Particle Swarm Optimization–Support Vector Machine (APSO-SVM) and Acceleration Artificial Bee Colony –Improved Transductive SVM (AABC-ITSVM) is introduced to handle the feature selection process over big data. However these methods are issues with computational complexity and accuracy of dataset. To avoid the above mentioned issues, in the proposed system, AABC-Artificial Neural Network (ANN) is proposed. This research contains modules are such as preprocessing, feature selection and classification. In preprocessing, k-Nearest Neighbor algorithm is applied which is used to handle the noise data efficiently. The size of the dataset is reduced significantly. Then these features are taken into feature selection process. In this research, AABC algorithm is used for performing feature selection. AABC optimization algorithm is used to select the important and relevant features from the preprocessed data. Then classification is done by using Artificial Neural Network (ANN) and it classifies more accurate classification results for the given large volume of dataset. ANN contains three layers are such as input, hidden and output layer. It is proposed to improve the time complexity by using neurons. The experimental result proves that the proposed system gives superior performance in terms of higher accuracy, recall, precision, f-measure, and lower time complexity by using AABC-ANN approach.Item FEATURE SELECTION USING MODIFIED PARALLEL SOCIAL SPIDER OPTIMIZATION ALGORITHM BASED ON SWARM INTELLIGENCE FOR HIGH DIMENSIONAL DATASETS(Dr. N.G.P Arts and Science College, 2018-08-09) S, Meera; B, Rosiline JeethaThe Social-Spider Optimization (SSO) is one among the recently developed swarm intelligence. It is impressed from the social behavior of spiders living in huge colonies. The spiders communicate among themselves to find prey and to find a suitable partner for mating. In this manuscript, a parallel version of this algorithm is modified and termed as MPSSO. Social spider optimization algorithm (SSO) was applied to develop a clustering algorithm since it does not favor the premature convergence or damage the exploration–exploitation balance. This scheme requirements a large number of iteration to achieve the desired convergence. So, a new parallel SSO was proposed via making the sequential movements of dominate male, non-dominate male and female spiders to parallel, in this manner decreasing the computational complexity of the algorithm. This approach was based on uniform distribution. The primary objective of the modified parallel SSO (MPSSO) scheme based on beta distribution and natural gradient (NG) local search for improving the performance of PSSO algorithm. The proposed approach is achieved a better exploration/exploitation trade-off while applied to optimization issues in the continuous domain. In this approach, a beta distribution is applied to tune the control parameters and a local search process is improved by the usage of the natural gradient.Item COMPARATIVE STUDY ON SWARM SEARCH FEATURE SELECTION FOR BIG DATA STREAM MINING(PSGR Krishnammal College for Women, 2017-02-04) S, MeeraIn the modern world there is huge development in the field of networking technology which handles huge data at a time. This data can be structured, semi structured or unstructured. To perform efficient mining of valuable information from such type of data the big data technology is gaining importance nowadays. Data mining application is been used in public and private sectors of industry because of its advantage over conventional networking technology to analyze large real time data. Data mining mainly relies on 3 V’s namely, Volume, Varity and Velocity of processing data. Volume refers to the huge amount of data it collects, Velocity refers to the speed at which it process the data and Variety defines that multi-dimensional data which can be numbers, dates, strings, geospatial data, 3D data, audio files, video files, social files, etc. These data which is stored in big data will be from different source at different rate and of different type; hence it will not be synchronized. This is one of the biggest challenges in working with big data. Second challenge is related to mining the valuable and relevant information from such data adhering to 3rd V i.e. Velocity. Speed is highly important as it is associated with cost of processing. On the other hand, mining through the high dimensional data the search space from which an optimal feature subset is determined and it is enhanced in size, guiding to a difficult stipulate in computation. With respect to handle the troubles, the research work is generally based on the high-dimensionality and streaming structure of data feeds in big data, a new inconsequential feature selection methodology that can be used to identify the feature selection methods in the big data. Some of the research work illustrates the different kinds of optimization methods for data stream mining would lead to tremendous changes in big data. This research work is focused on discussing various research methods that focus on finding the efficient feature selection methods which is used to avoid main challenges and produce optimal solutions. The previous methods are described with their advantages and disadvantages, consequently that the additional research works can be focused more. The tentative experiments were on the entire research works in Mat lab simulation surroundings and it is differentiated with everyone to identify the good methodologies beneath the different performance measures.