Search

The rough set theory is a powerful numerical model used to handle the impreciseness and ambiguity of data. Many existing multigranulation rough set models were derived from the multigranulation decision-theoretic rough set framework. The multigranulation rough set theory is very desirable in many practical applications such as high-dimensional knowledge discovery, distributional information systems, and multisource data processing. So far research works were carried out only for multigranulation rough sets in extraction, selection of features, reduction of data, decision rules, and pattern extraction. The proposed approach mainly focuses on anomaly detection in qualitative data with multiple granules. The approximations of the dataset will be derived through multiequivalence relation, and then, the rough set-based entropy measure with weighted density method is applied on every object and attribute. For detecting outliers, threshold value fixation is performed based on the estimated weight. The performance of the algorithm is evaluated and compared with existing outlier detection algorithms. Datasets such as breast cancer, chess, and car evaluation have been taken from the UCI repository to prove its efficiency and performance.

Data mining is an emerging technology where researchers explore innovative ideas in different domains, particularly detecting anomalies. Instances in the dataset which considerably deviate from others by their common patterns are known as anomalies. The state of being ambiguous and not affording certainty of data exists in this world of nature. Rough Set Theory is a proven methodology which deals with ambiguity and uncertainty of data. Research works that have been done until this point were focused on numeric or categorical type, which fails when the attributes are mixed type. By using fuzzy proximity and ordering relations, the numerical data has been converted to categorical data. This article presented an idea for detecting outliers in mixed data where the weighted density values of attributes and objects are calculated. The proposed approach has been compared with existing outlier detection methods by taking the hiring dataset as an example and benchmarked with Harvard dataverse datasets to prove its efficiency and performance.

Nowadays, the number of internet users is increasing vastly. Hence, predicting the number of channels for user’s communication is a major task. Therefore, scheduling of all traffic flows within the communication services in Long Term Evolution (LTE) scheduling is done by verifying the channel information and user availability in the network. In view of that, this paper proposes a novel feature extraction and classification method to evaluate the user availability status and the channel information for the betterment of communication within the LTE network. For preprocessing, we present a Fast Independent Component Analysis (FICA) method which incorporates dimensionality reduction in the given input feature data during the training course. In this work, feature extraction algorithm is used to extract the network feature and its degree of angle by means of distance between the features set. Subsequently, analyzes the extracted categories of network architecture by the analyzing weight value of that attribute based on the responses. As a result both the feature identification and framing of network structure increases the performance of data mining analysis. Hence, proposing a novel optimization technique like Particle Connected Cuckoo Search (PCCS) optimization algorithm for selecting the best features in the training feature set. Based on the extracted features, the classification method is performed in order to predict the network category by using Multilevel Neural Network (MLNN) classification technique. At this point, a novel kernel model for classification process is incorporated to reduce the time complexity. After that, the information is passed to the LTE scheduling system for providing enhanced communication. The comparative analysis between proposed techniques with the existing methods such as naive Bayes, stacking C, DTNB, random forest, J48, Ridor, decision table, zero R, grading in terms of accuracy, RMSE, CCI, ICI assures that the effectiveness of the proposed MLNN classification method.

Clustering algorithms are designed to deal with the datasets from a wide range of real-world applications based on the task. However, there are some situations where the clusters fail to maintain their performance due to mixed datasets problems. An active research area that is severely affected by these problems is the heart disease dataset. Deep learning methods, the state-of-the-art classifiers, with better learning procedures and computational resources, can fill these gaps. To improve the robustness of clusters, we propose a Constraint-Based Deep Convolutional Generative Adversarial Network (CB-DCGANs) framework for generating simulated data to augment the training set to improve the performance of the clustering algorithm. We evaluated the performance of Discriminatively Boosted Clustering (DBC) in detecting the clusters from given datasets. This study shows that using generative adversarial networks for clustering augmentation can significantly improve performance, especially in real-life applications.

In this modern era of computing, information technology revolution has brought drastic changes in the way data are collected for knowledge mining. The data thus collected are of value when it provides relevant knowledge pertaining to the interest of an organization. Therefore, the real challenge lies in converting high dimensional data into knowledge and to use it for the development of the organization. The data that is collected are generally released on internet for research purposes after hiding sensitive information in it and therefore, privacy preservation becomes the key factor for any organization to safeguard the internal data and also the sensitive information. Much research has been carried out in this regard and one of the earliest is the removal of identifiers. However, the chances of re-identification are very high. Techniques like k-anonymity and l-diversity helps in making the records unidentifiable in their own way, but, these techniques are not fully shielded against attacks. In order to overcome the drawbacks of these techniques, we have proposed improved versions of anonymization algorithms. The result analysis show that the proposed algorithms are better when compared to existing algorithms. Index Terms—K-anonymity, l-diversity, Data Publication, Anonymization, Privacy preservation, Generalization, Suppression, Datafly algorithm.

With the increasing size of data-sets in application areas like bio-medical, hospitals, information systems, scientific data processing and predictions, finance analytics, communications, retail and marketing, it is becoming increasingly important to execute data mining tasks in parallel. At the same time, technological advancements have made shared memoryparallel computation machines commonly available to various organizations and individuals. This paper analyzes a hierarchical clustering algorithm named chameleon clustering which is based on dynamic modeling and we propose a parallel algorithm for the same. The algorithm utilizes the concept of parallel processors available and hence reduces the time to generate final clusters. General Terms Enhancement, Utilization, clustering, similarity, Algorithms et. al.

Data Mining analyses the stocked data and helps in foretelling the future trends. There are different techniques by which data can be mined. These different techniques reveal different types of hidden knowledge. Using the right procedure of technique result specific patterns emerge. Ontology is a specification of conceptualization. It is a description of concepts and relationships that can exist for an agent or a community of agents. To make software more user-friendly, ontology could be used to explain both the technical and domain details. In the process of analyzing a data certain important details cannot be revealed, therefore security is the most important feature dealt in all technologies and work places. Data mining and Ontology techniques when integrated would capitulate an efficient system capable of selecting the appropriate algorithm for a data mining technique and privacy preserving techniques also by exploring the domain knowledge using ontology.