Probability based document clustering and image clustering using content-based image retrieval

Karthikeyan, M.; Aruna, P.

doi:10.1016/j.asoc.2012.09.013

Cited by 28 publications

(16 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However their work neither covered how to deal with multiple defined labels for a single document, nor included how to adapt the model to the latest information. Karthikeyan and Aruna [18] proposed a probabilitybased topic-oriented method with semi-supervision for text and image clustering and classification, and the result demonstrated improved accuracy in comparison to the DBSCAN and K-means algorithms. Their approach of probability calculation by using counts of words within documents is similar to the fundamentals of LDA, however, they did not address incremental adjustment of the model to make it adapt to the latest information.…”

Section: Introductionmentioning

confidence: 89%

Incremental learning with partial-supervision based on hierarchical Dirichlet process and the application for document classification

Wang

Alrubaie

2015

Applied Soft Computing

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 89%

Incremental learning with partial-supervision based on hierarchical Dirichlet process and the application for document classification

Wang

Alrubaie

2015

Applied Soft Computing

View full text Add to dashboard Cite

“…In modern era, many algorithms are there for document clustering process. Karthikeyan and Aruna [10] developed a method for document clustering on the basis of probability based semi-supervised document clustering. In this method, documents are clustered with respect to the user's need and this need is represented by multiple-attribute topic structure, it uses some topic annotations for each document for calculating similarity values between the documents.…”

Section: Related Workmentioning

confidence: 99%

Probabilistic Ranking of Documents Using Vectors in Information Retrieval

Saini

Singh

2014

Computational Intelligence in Data Mining - Volume 1

View full text Add to dashboard Cite

On the web, electronic form of information is increasing exponentially with the passage of past few years. Also, this advancement creates its own uncertainties. The overload information result is progressive while finding the relevant data with a chance of HIT or Miss Exposure. For improving this, Information Retrieval Ranking, Tokenization and Clustering techniques are suggestive as probable solutions. In this paper, Probabilistic Ranking using Vectors (PRUV) algorithm is proposed, in which tokenization and Clustering of a given documents are used to create more precisely and efficient rank gratify user's information need to execute sharply reduced search, is believed to be a part of IR. Tokenization involves pre-processing of the given documents and generates its respective tokens and then based on probability score cluster are created. Performance of some of existing clustering techniques (K-Means and DB-Scan) is compared with proposed algorithm PRUV, using various parameters, e.g. Time, Accuracy and Number of Tokens Generated.

show abstract

“…There are many other clustering techniques available and each of them may result in a different grouping of datasets and may not be effective for each type of dataset. Moreover, these clustering techniques may not scale to large datasets because of their computational time [24] and low accuracy. Therefore, many cluster analysis techniques are being developed for specific practical problems [25,26], such as finding classes of genes that have similar functions, grouping information on the Internet for different specific queries, and clustering biomedical data [27,28].…”

Section: Cluster Analysismentioning

confidence: 99%

“…To this end, many studies are specific to cluster analysis on image datasets [29], [30], focusing on the object shapes, joints, and localization [31]; colors [1,32]; and texture or image segmentation [1,33] through which image regions are discriminated. Likewise, Karthikeyan [24] studies image clustering using content-based image retrieval (CBIR) on whole images. He utilizes hue saturation value (HSV) color space and color histograms to acquire and cluster color similarities.…”

Section: Cluster Analysismentioning

confidence: 99%

Novel hybrid object-based non-parametric clustering approach for grouping similar objects in specific visual domains

Kuru

Khan

2018

Applied Soft Computing

View full text Add to dashboard Cite

Current widely employed clustering approaches may not yield satisfactory results with regard to the characteristics and distribution of datasets and number of clusters to be sought, especially for visual domains in multidimensional space. This study establishes a novel clustering methodology using a pairwise similarity matrix, Clustering Visual Objects in Pairwise Similarity Matrix (CVOIPSM), for grouping similar objects in specific visual domains. A dimensionality reduction and feature extraction technique, along with a distance measuring method and a newly established algorithm, Clustering in Pairwise Similarity Matrix (CIPSM), are combined to develop the CVOIPSM methodology. CIPSM utilizes both Rk-means and an agglomerative, contractible, expandable (ACE) technique to calculate a membership degree based on maximizing inter-class similarity and minimizing intra-class similarity. CVOIPSM has been tested on several datasets, with average success rates on downsized subsamples between 87.5% and 97.75% and between 81% and 87% on the larger datasets. The difference in the success rates for small and large datasets is not statistically significant (p>0.01). Moreover, this method automatically determines the likely number of clusters without any user dictation. The empirical results and the statistical significance test on these results ensure that CVOIPSM performs effectively and efficiently on specific visual domains, disclosing the interrelated patterns of similarities among objects.

show abstract

Probability based document clustering and image clustering using content-based image retrieval

Cited by 28 publications

References 16 publications

Incremental learning with partial-supervision based on hierarchical Dirichlet process and the application for document classification

Incremental learning with partial-supervision based on hierarchical Dirichlet process and the application for document classification

Probabilistic Ranking of Documents Using Vectors in Information Retrieval

Novel hybrid object-based non-parametric clustering approach for grouping similar objects in specific visual domains

Contact Info

Product

Resources

About