Download PDF by Sankar K Pal; Pabitra Mitra : Pattern recognition algorithms for data mining :

By Sankar K Pal; Pabitra Mitra

ISBN-10: 1584884576

ISBN-13: 9781584884576

Show description

Read Online or Download Pattern recognition algorithms for data mining : scalability, knowledge discovery and soft granular computing / [...] XA-GB PDF

Similar algorithms books

Mastering Algorithms with C by Kyle Loudon PDF

There are numerous books on facts buildings and algorithms, together with a few with worthy libraries of C services. gaining knowledge of Algorithms with C provides you with a special blend of theoretical heritage and dealing code. With strong options for daily programming projects, this publication avoids the summary form of such a lot vintage facts buildings and algorithms texts, yet nonetheless presents all the details you must comprehend the aim and use of universal programming recommendations.

Read e-book online Computer Graphics and Geometric Modeling: Implementation and PDF

Most likely the main accomplished evaluation of special effects as visible within the context of geometric modelling, this quantity paintings covers implementation and thought in a radical and systematic style. special effects and Geometric Modelling: Implementation and Algorithms, covers the pc pics a part of the sector of geometric modelling and contains all of the commonplace special effects subject matters.

Extra resources for Pattern recognition algorithms for data mining : scalability, knowledge discovery and soft granular computing / [...] XA-GB

Sample text

Let ng denote the © 2004 by Taylor & Francis Group, LLC Multiscale Data Condensation 33 current number of samples in GRABBAG whenever Step 1 of the algorithm is entered. 1. Use the k-NN rule with the current contents of STORE to classify the ith point from GRABBAG. If classified correctly the point is returned to GRABBAG; otherwise, it is placed in STORE. Repeat this operation for i = 1, 2, . . , ng . 2. If one complete pass is made through Step 1 with no transfer from GRABBAG to STORE, or the GRABBAG is exhausted then terminate; else go to Step 1.

Clustering is used in several exploratory data analysis tasks, customer retention and management, and web mining. The clustering problem has been studied in many fields, including statistics, machine learning and pattern recognition. However, large data considerations were absent in these approaches. Recently, several new algorithms with greater emphasis on scalability have been developed, including those based on summarized cluster representation called cluster feature (Birch [291], ScaleKM [29]), sampling (CURE [84]) and density joins (DBSCAN [61]).

Each level of the tree represents a partition of the feature space at a particular scale of detail. Prediction for a query point is performed using blocks from different scales; finer scale blocks are used for points close to the query and cruder scale blocks for those far from the query. However, the blocks are constructed by simple median splitting algorithms which do not directly consider the density function underlying the data. We describe in this chapter a density-based multiresolution data reduction algorithm [165] that uses discs of adaptive radii for both density estimation and sample pruning.

Download PDF sample

Pattern recognition algorithms for data mining : scalability, knowledge discovery and soft granular computing / [...] XA-GB by Sankar K Pal; Pabitra Mitra


by Joseph
4.0

Rated 4.49 of 5 – based on 37 votes