Indexing in Big Data Mining and Analytics

Abdullahi, Ali Usman; Ahmad, Rohiza; Zakaria, Nordin

doi:10.1007/978-3-030-66288-2_5

Cited by 3 publications

(2 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A variety of applications can be found in big data, such as social networks and healthcare. A comprehensive study of indexing technologies for big data was recently done by Abdullahi et al [2], which indicated that handling big data successfully and efficiently is challenging. Due to an exponential increase in the variety of cyberattacks, this challenge has played a pivotal role not just in indexing and membership queries, but likewise in big data security analytics [3].…”

Section: Introductionmentioning

confidence: 99%

The Impact of Check Bits on the Performance of Bloom Filter

Khan¹,

Qamar²,

Alsuhibany³

et al. 2022

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

Bloom filter (BF) is a space-and-time efficient probabilistic technique that helps answer membership queries. However, BF faces several issues. The problems with traditional BF are generally two. Firstly, a large number of false positives can return wrong content when the data is queried. Secondly, the large size of BF is a bottleneck in the speed of querying and thus uses large memory. In order to solve the above two issues, in this article, we propose the check bits concept. From the implementation perspective, in the check bits approach, before saving the content value in the BF, we obtain the binary representation of the content value. Then, we take some bits of the content value, we call these the check bits. These bits are stored in a separate array such that they point to the same location as the BF. Finally, the content value (data) is stored in the BF based on the hash function values. Before retrieval of data from BF, the reverse process of the steps ensures that even if the same hash functions output has been generated for the content, the check bits make sure that the retrieval does not depend on the hash output alone. This thus helps in the reduction of false positives. In the experimental evaluation, we are able to reduce more than 50% of false positives. In our proposed approach, the false positives can still occur, however, false positives can only occur if the hash functions and check bits generate the same value for a particular content. The chances of such scenarios are less, therefore, we get a reduction of approximately more than 50% false positives in all cases. We believe that the proposed approach adds to the state of the art and opens new directions as such.

show abstract

Section: Introductionmentioning

confidence: 99%

The Impact of Check Bits on the Performance of Bloom Filter

Khan¹,

Qamar²,

Alsuhibany³

et al. 2022

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

show abstract

“…Big data has several applications, such as in social networks and healthcare. A recent survey by the authors of reference [ 2 ] indicates that proficiently handling big data is competitive and challenging. This challenge centers mainly on big data security analytics (BDSA) because of a rapid increase in the type of cyber-attacks [ 3 ].…”

Section: Introductionmentioning

confidence: 99%

Performance Analysis of Bloom Filter for Big Data Analytics

Alsuhibany

Alsuhaibani

Khan

et al. 2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

The rapid rise of data value, such as social media and mobile applications, results in large volumes of data, which is what the term “big data” refers to. The increased rate of data growth makes handling big data very challenging. Despite a Bloom filter (BF) technique having previously been proposed as a space-and-time efficient probabilistic method, this proposal has not yet been evaluated in terms of big data. This study, thus, evaluates the BF technique by conducting an experimental study with a large amount of data. The results revealed that BF overcomes the efficiency not present in the space-and-time of indexing and examining big data. Moreover, to address the increase of false-positive rate in using BF with big data, a novel false-positive rate reduction approach is proposed in this paper. The initial experimental results of evaluating this method are very promising. The novel approach helped to reduce the false-positive rate by more than 70%.

show abstract