An improved K-nearest-neighbor algorithm for text categorization

Jiang, Shengyi; Pang, Guansong; Wu, Meiling; Kuang, Limin

doi:10.1016/j.eswa.2011.08.040

Cited by 239 publications

(97 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The main mission of TC is to allocate a Boolean value to every pair in D [11]. Figure 1 shows that, various categories are found in the document domain D and set C. D includes three different types of documents, namely '#', '$' and '@,'after categorization.…”

Section: Introductionmentioning

confidence: 99%

Categorization of ‘Holy Quran-Tafseer’ using K-Nearest Neighbor Algorithm

Hassan¹,

Mohammad²,

Alwan³

2015

IJCA

View full text Add to dashboard Cite

Text categorization, TC, is a process of labeling natural language texts with one or several categories from a predefined set. TC is a supervised learning where the set of categories and examples of documents belonging to those categories is given. The task of automatic TC is assigned an electronic document to several categories, based on a training set of labeled documents. The research objectives are, to formulate a K-Nearest Neighbor (KNN) algorithm for the automatic and suitable classification of any Holy Quran Tafseer segment; to identify relevant categories of Holy Quran Tafseer in the form of number classes; and to retrieve, Tafseer of verses of the Holy Quran in Malay language. Hence, this research aims to automatically categorize the Tafseer of verses of Holy Quran using the KNN algorithm as a technique to solve text categorization. This research has been designed to classify different verses in the Holy Quran. The first phase is to pre-process the Arabic text and then change the word in Arabic to Malay word. After that, categorize classes based on the cosine similarity between a test document and specific training documents. The majority of the same kind of nearest neighbors decides the category of the test sample and calculates precision and recall for a collection of documents. The result shows the outperform of TC using the KNN algorithm is one of the best algorithm for categorization Tafseer of Holy Quran. Furthermore, this study contributes in building a classifier to Tafseer Al-Quran in Malay language.

show abstract

Section: Introductionmentioning

confidence: 99%

Categorization of ‘Holy Quran-Tafseer’ using K-Nearest Neighbor Algorithm

Hassan¹,

Mohammad²,

Alwan³

2015

IJCA

View full text Add to dashboard Cite

show abstract

“…It has been widely used in pattern recognition and image processing applications, such as text categorization [15], gene classification [16], content-based image retrieval [17], image compression [18], and so on.…”

Section: K-nearest Neighbor (Knn) Algorithmmentioning

confidence: 99%

“…A reconfigurable associative memory (RASM) concept is developed to reconfigure the reference storage and vector dimensionality with vectorcomponent (word) parallelism by programmable switches (PS). RASM is a complementary solution to the DEC method [14][15][16][17][18][19]. Assume that the RASM has elements arranged in R rows and M columns, which contain SRAM cells for p vector components, p vectordistance computing units (DCUs) and one DEU.…”

Section: Programmable Switches For Reconfiguring Reference Storage Anmentioning

confidence: 99%

K-Nearest Neighbor Associative Memory with Reconfigurable Word-Parallel Architecture

An¹,

Mihara²,

Yamasaki³

et al. 2016

JSTS:Journal of Semiconductor Technology and Science

View full text Add to dashboard Cite

Abstract-IC-implementations provide high performance for solving the high computational cost of pattern matching but have relative low flexibility for satisfying different applications. In this paper, we report an associative memory architecture for k nearest neighbor (KNN) search, which is one of the most basic algorithms in pattern matching. The designed architecture features reconfigurable vectorcomponent parallelism enabled by programmable switching circuits between vector components, and a dedicated majority vote circuit. In addition, the main time-consuming part of KNN is solved by a clock mapping concept based weighted frequency dividers that drastically reduce the in principle exponential increase of the worst-case search-clock number with the bit width of vector components to only a linear increase. A test chip in 180 nm CMOS technology, which has 32 rows, 8 parallel 8-bit vector-components in each row, consumes altogether in peak 61.4 mW and only 11.9 mW for nearest squared Euclidean distance search (at 45.58 MHz and 1.8 V).

show abstract

“…The KNN method [32] is extended to cluster the documents to construct the personalized knowledge map. The key steps in the KNN method are the determination of the similarity between the cluster and the document.…”

Section: Personalized Document Clusteringmentioning

confidence: 99%

“…The detailed processes of the document clustering based on the KNN include the following steps [32]. 1) Create a new empty cluster and read a document as the centroid of the cluster.…”

Section: B Text Clusteringmentioning

confidence: 99%

An Approach to the Construction of Personalized Knowledge Map Based on Collaborative Tagging

Li¹,

Yuan²,

Xiong³

2015

IJKE

View full text Add to dashboard Cite

Abstract-Knowledge is the strategic resource for the organization. Knowledge map is an important tool for knowledge sharing. Providing the personalized knowledge map based on the preference of users can ease the burden of learning the knowledge map and facilitate the finding of the required knowledge. The tagging to documents reflects the user's preference of classification. In the paper, the approach to the personalized knowledge map construction based on the collaborative tagging is proposed. Firstly, the weight of the tag in documents is identified. Secondly, the similarity of users on the preference of classification is defined and then users that have the similar classification preference are identified to expand the current user's preference of personalized classification. Then the text vector of document and text similarity is identified. Afterwards, the knowledge is clustered according to both the personalized classification similarity and text similarity. Finally, the topics of each cluster are identified. In the topic identification, both the weight of the term in the text and the weight of the term in the tags are integrated. The experiment shows that proposed method is feasible and performs well.Index Terms-Knowledge map, knowledge management systems, personalized knowledge map.

show abstract

An improved K-nearest-neighbor algorithm for text categorization

Cited by 239 publications

References 21 publications

Categorization of ‘Holy Quran-Tafseer’ using K-Nearest Neighbor Algorithm

Categorization of ‘Holy Quran-Tafseer’ using K-Nearest Neighbor Algorithm

K-Nearest Neighbor Associative Memory with Reconfigurable Word-Parallel Architecture

An Approach to the Construction of Personalized Knowledge Map Based on Collaborative Tagging

Contact Info

Product

Resources

About