Experiments with Semantic Similarity Measures Based on LDA and LSA

Niraula, Nobal B.; Banjade, Rajendra; Ştefănescu, Dan; Rus, Vasile

doi:10.1007/978-3-642-39593-2_17

Cited by 22 publications

(9 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Latent semantic analysis (LSA) given by Niraula and Banjade (2013) and LDA proposed by Vidhya and Aghila (2010), are two popular mathematical approaches to modelling textual data. Questions posed by algorithm developers and data analysts working with LSA and LDA models motivated to How closely do LSAs concepts correspond to LDAs topics?…”

Section: Method: Ldamentioning

confidence: 99%

A study to enhance candidate screening process using similarity analysis

Ujlayan

Sharma

2019

IJBDA

View full text Add to dashboard Cite

Recruitment process is always continuous and open positions mostly have short time to get filled thereby for any such open position, organisations generally have to deal with higher incremental costs. Many advanced recruitment processes/methodologies are directly associated with the risk of failure, loss of effort, loss of time and money. Every organisation in any industry always aims to hire the best suitable candidate to the open position with right skillsets in the shortest possible time and with the lowest costs. Therefore, there is a need to experiment and explore the existing approaches, which can help in reducing the time and the screening effort in initial stage of recruitment practices. Consequently this paper attempts to implement the similarity analysis approach for a random sample of resumes from IT sector in National Capital Region (NCR), India to provide the best match of candidates' profile as per the required job description. The latent Dirichlet allocation (LDA) is used to find the similarity between job description and candidate's profile. The study will help the IT industry in identifying and selecting the best matched candidates' profile based on the key features in job description.

show abstract

Section: Method: Ldamentioning

confidence: 99%

A study to enhance candidate screening process using similarity analysis

Ujlayan

Sharma

2019

IJBDA

View full text Add to dashboard Cite

show abstract

“…Recent studies have shown that similarity measures of features are more efficient when based on topic models techniques than they are based on bag of words and TF-IDF ( Xie & Xing, 2013 ). In this context, the semantic similarity between two documents was also investigated ( Niraula et al, 2013 ). The most related work to our context is probably the use of topic modeling features to improve the word sense disambiguation by Li & Suzuki (2021) and also the work in Pavlinek & Podgorelec (2017) in which they present features representation with a semi-supervised approach using self-training learning.…”

Section: Related Workmentioning

confidence: 99%

Topic2features: a novel framework to classify noisy and sparse textual data using LDA topic distributions

Wahid

Shi

Gao

et al. 2021

PeerJ Computer Science

View full text Add to dashboard Cite

In supervised machine learning, specifically in classification tasks, selecting and analyzing the feature vector to achieve better results is one of the most important tasks. Traditional methods such as comparing the features’ cosine similarity and exploring the datasets manually to check which feature vector is suitable is relatively time consuming. Many classification tasks failed to achieve better classification results because of poor feature vector selection and sparseness of data. In this paper, we proposed a novel framework, topic2features (T2F), to deal with short and sparse data using the topic distributions of hidden topics gathered from dataset and converting into feature vectors to build supervised classifier. For this we leveraged the unsupervised topic modelling LDA (latent dirichlet allocation) approach to retrieve the topic distributions employed in supervised learning algorithms. We made use of labelled data and topic distributions of hidden topics that were generated from that data. We explored how the representation based on topics affect the classification performance by applying supervised classification algorithms. Additionally, we did careful evaluation on two types of datasets and compared them with baseline approaches without topic distributions and other comparable methods. The results show that our framework performs significantly better in terms of classification performance compared to the baseline(without T2F) approaches and also yields improvement in terms of F1 score compared to other compared approaches.

show abstract

“…In the general English domain, semantic evaluation (SemEval) STS shared tasks have been organized annually from 2012 to 2017 [1][2][3][4][5][6], and STS benchmark datasets were developed for evaluation [6]. Previous work on STS often used machine learning models [7][8][9] such as support vector machine [10], random forest [11], convolutional neural networks [12], and recurrent neural networks [13] and topic modeling techniques [8] such as latent semantic analysis [14] and latent Dirichlet allocation [15]. Recently, deep learning models based on transformer architectures such as Bidirectional Encoder Representations from Transformers (BERT) [16], XLNet [17], and Robustly optimized BERT approach (RoBERTa) [18] have demonstrated state-of-the-art performances on the STS benchmark dataset [19] and remarkably outperformed the previous models.…”

Section: Introductionmentioning

confidence: 99%

Measurement of Semantic Textual Similarity in Clinical Texts: Comparison of Transformer-Based Models

Yang¹,

He²,

Zhang³

et al. 2020

JMIR Med Inform

View full text Add to dashboard Cite

Background Semantic textual similarity (STS) is one of the fundamental tasks in natural language processing (NLP). Many shared tasks and corpora for STS have been organized and curated in the general English domain; however, such resources are limited in the biomedical domain. In 2019, the National NLP Clinical Challenges (n2c2) challenge developed a comprehensive clinical STS dataset and organized a community effort to solicit state-of-the-art solutions for clinical STS. Objective This study presents our transformer-based clinical STS models developed during this challenge as well as new models we explored after the challenge. This project is part of the 2019 n2c2/Open Health NLP shared task on clinical STS. Methods In this study, we explored 3 transformer-based models for clinical STS: Bidirectional Encoder Representations from Transformers (BERT), XLNet, and Robustly optimized BERT approach (RoBERTa). We examined transformer models pretrained using both general English text and clinical text. We also explored using a general English STS dataset as a supplementary corpus in addition to the clinical training set developed in this challenge. Furthermore, we investigated various ensemble methods to combine different transformer models. Results Our best submission based on the XLNet model achieved the third-best performance (Pearson correlation of 0.8864) in this challenge. After the challenge, we further explored other transformer models and improved the performance to 0.9065 using a RoBERTa model, which outperformed the best-performing system developed in this challenge (Pearson correlation of 0.9010). Conclusions This study demonstrated the efficiency of utilizing transformer-based models to measure semantic similarity for clinical text. Our models can be applied to clinical applications such as clinical text deduplication and summarization.

show abstract

Experiments with Semantic Similarity Measures Based on LDA and LSA

Cited by 22 publications

References 12 publications

A study to enhance candidate screening process using similarity analysis

A study to enhance candidate screening process using similarity analysis

Topic2features: a novel framework to classify noisy and sparse textual data using LDA topic distributions

Measurement of Semantic Textual Similarity in Clinical Texts: Comparison of Transformer-Based Models

Contact Info

Product

Resources

About