Data Mining Clustering Algorithm Research Articles

An iterative algorithm for optimal variable weighting in K-means clustering

ABSTRACTThe K-means clustering method is a widely adopted clustering algorithm in data mining and pattern recognition, where the partitions are made by minimizing the total within group sum of squares based on a given set of variables. Weighted K-means clustering is an extension of the K-means method by assigning nonnegative weights to the set of variables. In this paper, we aim to obtain more meaningful and interpretable clusters by deriving the optimal variable weights for weighted K-means clustering. Specifically, we improve the weighted k-means clustering method by introducing a new algorithm to obtain the globally optimal variable weights based on the Karush-Kuhn-Tucker conditions. We present the mathematical formulation for the clustering problem, derive the structural properties of the optimal weights, and implement an recursive algorithm to calculate the optimal weights. Numerical examples on simulated and real data indicate that our method is superior in both clustering accuracy and computational efficiency.

An Attack Threat Effect Analysis Method Based on K-Means Evaluation

To take full advantage of the specified features of the attack dataset in network attack effect evaluation, maximize the efficiency of evaluation without losing its accuracy. This paper proposed a K-Means evaluation technique using dimensional entropy components, derived from changes in network entropy through attack period and the advantages of clustering algorithm in data mining. This method makes a pre-process of the attack dataset on the basis of network entropy, mapping it to a two-dimensional plane and utilize the output of pre-process as the input of clustering. Then establish a relation between the attack dataset and the effect category via an improved K-Means algorithm, thus achieving an explicit division of attack effect set and provide efficient evaluation result. The experimental results prove that the method can process attack dataset with high efficiency, as well as provide a visualized evaluation result by the form of cluster tree.

Enhanced web cache replacement policy based on data mining and RFSD scoring

By rapid growth of the Internet users and devices, the number of servers also increase simultaneously which causes the exponential increment of the internet traffic and static data. Handling these huge amounts of user requests and efficiently responding to them require high bandwidth links, powerful servers and robust equipment, which despite the availability of these requirements getting the full user satisfaction is extremely difficult and a tough challenge. In order to overcome the mentioned problem, the cache servers are being used as a suitable solution. The performance of web cache server directly depends on its replacement policies. Several cache replacement policies have been proposed in literature each having varied hit rate (HR) and byte hit rate (BHR) performances on different networks. The replacement policy proposed in this paper is a dynamic cache replacement policy which trains itself utilizing previous network logs and by exploiting the data mining clustering algorithm. Once the training step is completed, the proposed policy utilizes the normalization formulas to score each metric of the enquiries including recency, frequency, size and delay. Simulation results showed that the proposed policy has the optimum performance on different networks and it not only improved the performance of web cache server in term of HR and BHR, but also decreases the data retrieval time (Delay Ratio (DR)) of the cache servers.

English

English

Optimization of Physical Education Teaching Mode in Colleges based on Data Mining Clustering Algorithm

Optimization of Physical Education Teaching Mode in Colleges based on Data Mining Clustering Algorithm

Customized M-clustering Algorithm Comparison with Clustering Algorithms in Data Mining with the Case Study of Lead Generation Techniques

Objectives: Clustering algorithm is broadly used as spectral algorithm in social media, where a reference of contact is used and mined further for various combinations of suggested friends and lookups. This paper Identifies key lead generation techniques to be used in customer relationship management for sales marketing to decide data gathering. Also to define key merits and demerits of these techniques and to prepare a Matrix of comparison of these techniques to justify the data source and data set. M-Cluster algorithm is used for lead qualification i.e. training set preparation and data evaluation. Methods: Todefine a training set based on various attributes/fields in the data given for classification. This training set is used to run the data process and to produce expected result. This is assessed and accepted for definite data set processing or additional run for interim training data set preparation. Findings: This study is taken to customize clustering algorithm for data mining process in a customer relationship management field as the space of data is more and variant. Also proving the usability of customized clustering algorithm in data mining and the efficiency in processing mechanism compared to other methods used in current situation of data mining in customer relationship management is the part of this study. The customized algorithm developed as part of this study considers the two major areas of data mining using clusters. Applications: The results produced tremendous trends that the clustered algorithm suits to any data mining process when scaling and data classification are diversified and less in control.

A Survey of Data Mining Clustering Algorithms

is a technique used in data mining that groups similar objects into one cluster, while dissimilar objects are grouped into different clusters. The clustering techniques can be categorized into partitioning methods, hierarchical methods, density-based methods and grid-based methods. The different partitioning methods studied here are k-means and k- medoids. The different hierarchical techniques studied here are BIRCH and CHAMELEON. The different grid-based techniques which are described are DBSCAN and DENCLUE. Lastly, the different techniques which are used in grid-based technique, like STING and CLIQUE are described. This paper aims to provide a brief overview and comparison of these different clustering algorithms and methods.

Collaborative Data Processing in WSN Using Voronoi Fuzzy Clustering

In this paper, developed a novel Voronoi Fuzzy Clustering (VF) algorithm for energy efficient collaborative data aggregation in wireless sensor network. VF algorithm is fusion of Voronoi diagram and modified Fuzzy C- Means with respect to distance and Quality of Service. Here throughput, delay time and delivery ratio are considered as QOS parameters. Once clustering of sensor nodes is completed then data management technique such as data aggregation or compression is done for further decision making in sink node. Data mining clustering algorithm reduces overall transmission of data from each sensor to the sink node thus energy spent by individual sensor node is minimized. The cluster heads collects all sensed information from their respective cluster members and performs data aggregation or compression before transmitting the data to the sink node. Finally, the simulations are performed and the results are analyzed within the simulation set up to determine performance of the proposed algorithm in the sensor network. Our proposed approach has achieved 60% efficiency when compare with the K means algorithm.

Spatial Clustering Analysis of K-Means Algorithm in the Classification of Bank Card Customers

K-means algorithm is a simple and efficient data mining clustering algorithm. For the current status of the bank card customer relationship management, based on data mining technology, design based on K-means clustering algorithm banking customer classification system. Data mining techniques can extract vast amounts of customer information data bank card implicit knowledge and spatial relationship model will represent the bank customers feature set of data objects automatically classified into each composed of clusters of similar objects, bank card customers in the banking system classification. This paper analyzes the existing spatial clustering methods summary and conclusion, based on the combined data bank card customers, according to the volatility of funds used to different customer groups, the use of K-means analysis to study characteristics of client groups, providing appropriate financial services.

Data Mining Clustering Algorithm Research and Application

In this paper, first introduce data mining knowledge then focuses on the clustering analysis algorithms, including classification clustering algorithm, and each classification typical cluster analysis algorithms, including the formal description of each algorithm as well as the advantages and disadvantages of each algorithm also has a more detailed description. Then carefully introduce data mining algorithm on the basis of cluster analysis. And using cohesion based clustering algorithm with DBSCAN algorithm and clustering in consumer spending in two-dimensional space, 2,000 data points for each area, and get a reasonable clustering results, resulting in hierarchical clustering results valuable information, so as to realize the practical application of the algorithm and clustering analysis theory combined.

Adaptive Initialization Method Based on Spatial Local Information fork-Means Algorithm

k-means algorithm is a widely used clustering algorithm in data mining and machine learning community. However, the initial guess of cluster centers affects the clustering result seriously, which means that improper initialization cannot lead to a desirous clustering result. How to choose suitable initial centers is an important research issue fork-means algorithm. In this paper, we propose an adaptive initialization framework based on spatial local information (AIF-SLI), which takes advantage of local density of data distribution. As it is difficult to estimate density correctly, we develop two approximate estimations: density byt-nearest neighborhoods (t-NN) and density byϵ-neighborhoods (ϵ-Ball), leading to two implements of the proposed framework. Our empirical study on more than 20 datasets shows promising performance of the proposed framework and denotes that it has several advantages: (1) can find the reasonable candidates of initial centers effectively; (2) it can reduce the iterations ofk-means’ methods significantly; (3) it is robust to outliers; and (4) it is easy to implement.

Clustering Algorithm in Data Mining Based on Web Log

The advantages of FCM algorithm are that it is mainly applied in point data cluster and can't directly process relational data, for which the paper proposes a clustering algorithm in data mining based on web log . Firstly, the paper improves FCM algorithm which makes it can process relational data, and makes robustness improvement on the algorithm. Then, the traditional FCM algorithm needs to determine in advance on the basis without prior knowledge, for which the paper introduces competition agglomerative algorithm and makes it combine with FCM algorithm, which generates CA-FCM algorithm making it can automatically determine category number of the best classification. The experiments show that mining results of CA-FCM algorithm is close to the mining results of FCM algorithm, and the performance of CA-FCM algorithm is better than that of FCM algorithm when the amount of users access to session is not too much.

Performance Evaluation of Data Mining Clustering Algorithm in WEKA

Data mining is a computerized technology that uses complicated algorithms to find relationships and trends in large data bases, real or perceived, previously unknown to the retailer, to promote decision support.., data mining is touted to be one of the widespread recognition of the potential for analysis of past transaction data to improve the quality of future business decisions. The purpose of this paper is to critique data mining technology in comparison with more familiar data mining algorithm in well known tool Weka for strategic decision making by small to medium size retailers. The context for this study includes current and future industry applications and practices for research performed in data mining applications within the retail sector.

Clustering Methods and Algorithms in Data Mining: Review

Clustering Methods and Algorithms in Data Mining: Review

A decision making process application for the slurry production in ceramics via fuzzy cluster and data mining

To increase productivity, companies are in search of techniques that enable them to make faster and more effective decisions. Data mining and fuzzy clustering algorithms can serve for this purpose. This paper models the decision making process of a ceramics production company using a fuzzy clustering algorithm and data mining. Factors that affect the quality of slurry are measured over time. Using this data, a fuzzy clustering algorithm assigns the degrees of memberships of the slurry for the different quality clusters. An expert can decide on acceptance or rejection of slurry based on calculated degrees of memberships. In addition, by using data mining techniques we generated some rules that provide the optimum conditions for acceptance of the slurry.

An Algorithm for Detecting Community Structure of Complex Networks based on Clustering

There are considerable interest in algorithms for detecting community structure, which is fundamental for analyzing the relationship between structure and function in complex networks. In this paper, after the introduction of some traditional approaches for detecting community structure and data mining clustering algorithms, we propose Mapping Vertex into Vector(MVV) algorithm, which can convert network nodes to vector, based on the algorithm, we propose K-means algorithm for detecting community structure based on MVV, and use fuzzy c-means to deal with overlapping community. Finally, experiments show that the algorithm presented in this paper is of high accuracy with good performance.

A Survey of Partition based Clustering Algorithms in Data Mining: An Experimental Approach

A Survey of Partition based Clustering Algorithms in Data Mining: An Experimental Approach

Fuzzy based clustering algorithm for privacy preserving data mining

Sharing of data among multiple organisations is required in many situations. The shared data may contain sensitive information about individuals which if shared may lead to privacy breach. Thus, maintaining the individual privacy is a great challenge. In order to overcome the challenges involved in data mining, when data needs to be shared, privacy preserving data mining (PPDM) has evolved as a solution. The objective of PPDM is to have the interesting knowledge mined from the data at the same time to maintain the individual privacy. This paper addresses the problem of PPDM by transforming the attributes to fuzzy attributes. Thus, the individual privacy is also maintained, as one cannot predict the exact value, at the same time, better accuracy of mining results is achieved. ID3 and Naive Bayes classification algorithms over three different datasets are used in the experiments to show the effectiveness of the approach.

Hybrid Approach of Data Mining clustering algorithms

Clustering is a way that classifies the raw data reasonably and searches the hidden patterns that may exist in datasets. It is a process of grouping data objects into disjoint clusters so that data in the same cluster are similar, and data belonging to different cluster are differ. Many algorithms have been developed for clustering. In this paper we are reviewing different clustering algorithms like K-Means , HAC , SOM and their comparison by applying hybrid approach on different datasets. Keywords: Clustering Algorithms, Data Mininig, HAC, SOM, K-Means

A Kind of Hierarchical K-Means Web Log Clustering Algorithm

Clustering techniques are often used in Web log mining to analyze user’s interest on the web pages. Based on the analysis of advantages and disadvantages of the application of classic clustering algorithm in Web log data mining, the paper brought out a kind of hierarchical K-means Web log clustering algorithm, which integrated K-means clustering algorithm and cohesion-based hierarchical clustering algorithm and overcame shortcoming of high time complexity of hierarchical clustering algorithm. The clustering effect of the algorithm is better than K-means clustering and fit for clustering process of large amount data. The result analysis of practical Web log data clustering also proves the validity of the algorithm.