Survey of Text Mining II

Berry, Michael W.; Castellanos, Malú

doi:10.1007/978-1-84800-046-9

Cited by 57 publications

(5 citation statements)

References 98 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have also tested the modified Lee-Seung method of Finesso and Spreij [15,51], but it is only a scaled version of the Lee-Seung method for KL divergence and gave nearly identical results which are omitted. All implementations are from Version 2.5 of Tensor Toolbox for MATLAB [4,3,2]; exact parameter settings are provided in Appendix E. We report the factor match score (FMS), a measure in [0, 1] of how close the computed solution is to the true solution. A value of 1 is ideal.…”

Section: Comparison Of Objective Functions For Sparse Count Datamentioning

confidence: 99%

“…In this paper, we consider the problem of multilinear modeling of sparse count data. For instance, we may consider data that encodes the number of papers published by each author at each conference per year for a given time frame [13], the number of packets sent from one IP address to another using a specific port [47], or to/from and term counts on emails [2]. Our goal is to develop a descriptive model of such data, along with appropriate algorithms and theory.…”

mentioning

confidence: 99%

See 1 more Smart Citation

On Tensors, Sparsity, and Nonnegative Factorizations

Chabrière¹,

Kolda²

2012

SIAM J. Matrix Anal. & Appl.

226

244

View full text Add to dashboard Cite

Abstract. Tensors have found application in a variety of fields, ranging from chemometrics to signal processing and beyond. In this paper, we consider the problem of multilinear modeling of sparse count data. Our goal is to develop a descriptive tensor factorization model of such data, along with appropriate algorithms and theory. To do so, we propose that the random variation is best described via a Poisson distribution, which better describes the zeros observed in the data as compared to the typical assumption of a Gaussian distribution. Under a Poisson assumption, we fit a model to observed data using the negative log-likelihood score. We present a new algorithm for Poisson tensor factorization called CANDECOMP-PARAFAC Alternating Poisson Regression (CP-APR) that is based on a majorization-minimization approach. It can be shown that CP-APR is a generalization of the Lee-Seung multiplicative updates. We show how to prevent the algorithm from converging to non-KKT points and prove convergence of CP-APR under mild conditions. We also explain how to implement CP-APR for large-scale sparse tensors and present results on several data sets, both real and simulated.

show abstract

Section: Comparison Of Objective Functions For Sparse Count Datamentioning

confidence: 99%

mentioning

confidence: 99%

On Tensors, Sparsity, and Nonnegative Factorizations

Chabrière¹,

Kolda²

2012

SIAM J. Matrix Anal. & Appl.

226

244

View full text Add to dashboard Cite

show abstract

“…In contrast, other text mining methods such as topic modelling and clustering can analyse groups of words together instead of counting them separately (Berry and Castellanos, 2008). In this way, these techniques can consider the variations in each word's meaning across different contexts.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Although conventional clustering methods overcome some of the limitations of term-frequency-based techniques, topic modelling presents several advantages over them. Clustering is based on the idea of classifying documents into different groups based on a specific similarity measure (Berry and Castellanos, 2008). With clustering, every document (reviews, in our case) appears only in one group (such as topics or attributes).…”

Section: Literature Reviewmentioning

confidence: 99%

Quality 4.0: big data analytics to explore service quality attributes and their relation to user sentiment in Airbnb reviews

Amat-Lefort

Barravecchia²,

Mastrogiacomo³

2022

IJQRM

View full text Add to dashboard Cite

PurposeQuality 4.0 is a new paradigm of quality management, which emphasises the need to adapt to recent technological innovations by updating traditional quality approaches. Amongst the most important factors for adopting Quality 4.0 is the leveraging of big data to collect insights and quality perceptions from clients. Therefore, user reviews have emerged as a valuable source of information, which can be analysed through machine learning procedures to uncover latent quality dimensions.Design/methodology/approachThis study applies a combination of text mining techniques to analyse Airbnb reviews, identifying service quality attributes and assessing their relation to the users' sentiment. More than two million reviews written by guests in four European cities are analysed. First, topic modelling is applied to find the quality attributes mentioned by reviewers. Then, sentiment analysis is used to assess the positiveness/negativeness of the users' feedback.FindingsA total of 37 quality attributes are identified. Four of them show a significant positive relation to the guest's sentiment: apartment views, host tips and advice, location and host friendliness. On the other hand, the following attributes are negatively correlated with user sentiment: sleep disturbance, website responsiveness, thermal management and hygiene issues.Originality/valueThis paper provides a practical example of how Quality 4.0 can be implemented, proposing a data-driven methodology to extract service quality attributes from user-generated content. Additionally, several attributes that had not appeared in existing Airbnb studies are identified, which can serve as a reference to extend previous quality assessment scales.

show abstract

“…Therefore we have to carefully define the experimentation environment [12] and the dataset that we use in our tests. tion we can show in Figure 4.…”

Section: Development Of Evolutionary System For Document Clusteringmentioning

confidence: 99%

Document Clustering with Evolutionary Systems through Straight-Line Programs “<i>slp</i>”

Sequera¹,

Díez²,

Sotos³

2012

JILSA

View full text Add to dashboard Cite

In this paper, we show a clustering method supported on evolutionary algorithms with the paradigm of linear genetic programming. "The Straight-Line Programs (slp)", which uses a data structure which will be useful to represent collections of documents. This data structure can be seen as a linear representation of programs, as well as representations in the form of graphs. It has been used as a theoretical model in Computer Algebra, and our purpose is to reuse it in a completely different context. In this case, we apply it to the field of grouping library collections through evolutionary algorithms. We show its efficiency with experimental data we got from traditional library collections.

show abstract

Survey of Text Mining II

Cited by 57 publications

References 98 publications

On Tensors, Sparsity, and Nonnegative Factorizations

On Tensors, Sparsity, and Nonnegative Factorizations

Quality 4.0: big data analytics to explore service quality attributes and their relation to user sentiment in Airbnb reviews

Document Clustering with Evolutionary Systems through Straight-Line Programs “<i>slp</i>”

Contact Info

Product

Resources

About

Survey of Text Mining II

Cited by 57 publications

References 98 publications

On Tensors, Sparsity, and Nonnegative Factorizations

On Tensors, Sparsity, and Nonnegative Factorizations

Quality 4.0: big data analytics to explore service quality attributes and their relation to user sentiment in Airbnb reviews

Document Clustering with Evolutionary Systems through Straight-Line Programs “&lt;i&gt;slp&lt;/i&gt;”

Contact Info

Product

Resources

About

Document Clustering with Evolutionary Systems through Straight-Line Programs “<i>slp</i>”