Semi supervised approach towards subspace clustering

Harikumar, Sandhya; A, Akhil

doi:10.3233/jifs-169456

Cited by 7 publications

(10 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For subspace clustering, we adapt an iterative approach of IGSC [19]. The outline of IGSC is shown in Algorithm 2.…”

Section: Algorithmsmentioning

confidence: 99%

“…Various types of subspace clustering methods exist in the literature such as CLIQUE [13], SUBCLU [14], and MAFIA [15] that finds clusters in all subspaces. There are other types of subspace clustering methods called projective clustering such as PROCLUS [16,17], FIND-IT [18], IGSC [19] and δ -CLUSTERS [20] that find clusters in axis-parallel subspaces. We follow a previous work where information gain-based semi-supervised-subspace clustering (IGSC) is suggested [19] to model the distributions of high-dimensional data with relevant features.…”

Section: Introductionmentioning

confidence: 99%

“…There are other types of subspace clustering methods called projective clustering such as PROCLUS [16,17], FIND-IT [18], IGSC [19] and δ -CLUSTERS [20] that find clusters in axis-parallel subspaces. We follow a previous work where information gain-based semi-supervised-subspace clustering (IGSC) is suggested [19] to model the distributions of high-dimensional data with relevant features. The details of IGSC algorithm are given in Section 4 and the algorithm is given in Algorithm 2.…”

Section: Introductionmentioning

confidence: 99%

“…Thus, given a query point, q , we aim to find the best distribution of points, S , with relevant features, that matches q , so as to minimize the distance between q and S to effectively retrieve appropriate label for q . For this, we adapt an existing subspace clustering called IGSC to obtain informative clusters, each with relevant dimensions using information gain [19]. Thereafter, we propose an improvised classification algorithm using a depth-based representation of subspace clusters, to improve scalability and accuracy of classification for high-dimensional data.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A depth-based nearest neighbor algorithm for high-dimensional data classification

Harikumar¹,

Savithri²,

Kaimal³

2019

Turk J Elec Eng & Comp Sci

Self Cite

View full text Add to dashboard Cite

Nearest neighbor algorithms like k-nearest neighbors (kNN) are fundamental supervised learning techniques to classify a query instance based on class labels of its neighbors. However, quite often, huge volumes of datasets are not fully labeled and the unknown probability distribution of the instances may be uneven. Moreover, kNN suffers from challenges like curse of dimensionality, setting the optimal number of neighbors, and scalability for high-dimensional data. To overcome these challenges, we propose an improvised approach of classification via depth representation of subspace clusters formed from high-dimensional data. We offer a consistent and principled approach to dynamically choose the nearest neighbors for classification of a query point by i) identifying structures and distributions of data; ii) extracting relevant features, and iii) deriving an optimum value of k depending on the structure of data by representing data using data depth function. We propose an improvised classification algorithm using a depth-based representation of clusters, to improve performance in terms of execution time and accuracy. Experimentation on real-world datasets reveals that proposed approach is at least two orders of magnitude faster for high-dimensional dataset and is at least as accurate as traditional kNN.

show abstract

“…For subspace clustering, we adapt an iterative approach of IGSC [19]. The outline of IGSC is shown in Algorithm 2.…”

Section: Algorithmsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A depth-based nearest neighbor algorithm for high-dimensional data classification

Harikumar¹,

Savithri²,

Kaimal³

2019

Turk J Elec Eng & Comp Sci

Self Cite

View full text Add to dashboard Cite

show abstract

“…In [35], the authors propose a multi-directional model for rule analysis and risk prediction of any disease based on input parameters. In [36], a top down semi-supervised subspace clustering is proposed to identify a subset of important attributes based on the known label for each data instance. In [37], the authors propose a training method that employs Deep Neural Networks (DNNs) to map the sensory data into a representation space aligned with human concepts.…”

mentioning

confidence: 99%

Soft computing and intelligent systems: Techniques and applications

Thampi

El-Alfy

Trajković

2018

IFS

View full text Add to dashboard Cite

This special issue is a collection of selected papers submitted to the Third International Symposium on Intelligent Systems Technologies and Applications (ISTA' 17) held in Manipal University, India, during September 13-16, 2017. These papers have been peer reviewed and accepted for presentation at the symposium. A second review has been conducted to improve content and presentation of the manuscripts published in this special issue. The 48 papers cover a wide range of powerful techniques and current applications of soft computing and intelligent systems.The International Symposium on Intelligent Systems Technologies and Applications aims to bring together researchers in related fields and provides a venue to explore and discuss various aspects of intelligent systems technologies and their applications. It provides excellent opportunities for the presentation and discussion of interesting new research results, which contributes to effective transfer of knowledge and dissemination of innovative ideas.Listed is a summary of the contributions of the papers pertaining to various application domains:Cybersecurity is an important concern for every organization. In this special issue, twelve papers * Corresponding author. Sabu M. Thampi, Indian Institute of Information Technology and Management-Kerala, Technopark Campus, Trivandrum-695581, Kerala State, India. E-mail: sabu.thampi@iiitmk.ac.in. deal with intrusion detection, integrity, and confidentiality. In [1], the authors describe an intrusion detection system for wireless mesh networks that identifies intrusion activities using a support vector machine classifier. The proposed method employs a hybrid genetic algorithm and a mutual-information technique for feature selection. In [2], a scheme is proposed to provide a secure outsourcing solution for the matrix-chain multiplication problem in cloud computing. For preserving data confidentiality, the client employs few linear transformation schemes. The proposed verification scheme helps to maintain the integrity of the computed result. In [3], the authors evaluate the efficiency of various deep learning methods to detect and classify domain names as either malicious or benign by automatically extracting the required features thus reducing the burden of manual feature engineering approaches. In [4], the effectiveness of recurrent neural network (RNN) and its variant long short-term memory (LSTM) is analyzed for Android malware detection of time-varying sequences of benign and malicious applications. This paper also provides a comparative analysis with other machine-learning classifiers. In [5], the authors propose a data-driven intrusion detection system for Industrial Internet of Things (IIoT) to hybridize both anomaly and specification-based schemes. In [6], an intelligent monitoring technique for smart grid

show abstract