Normalized Approach to Find Optimal Number of Topics in Latent Dirichlet Allocation (LDA)

Hasan, Mahedi; Rahman, Asheq; Karim, Razaul; Khan, Md. Saikat Islam; Islam, Md. Jahidul

doi:10.1007/978-981-33-4673-4_27

Cited by 49 publications

(16 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One area of contention in LDA is how many topics to use from a corpus (Zhao et al, 2015). Hasan et al (2021), in their research to examine the methods to find the optimal number of topics in LDA, established that a combination of perplexity score and coherence score works better than existing methods to determine the optimal number of topics. After reaching the optimal number of topics, the topics' relevance and meaningfulness must be assured.…”

Section: Topic Modellingmentioning

confidence: 99%

Decoding employee experiences during pandemic through online employee reviews: insights to organizations

Joshi

Sekar

Das

2023

View full text Add to dashboard Cite

PurposeThe purpose of this paper is to unearth various dimensions of employee experience (EX) and explore how pandemic impacted various EX factors using online employee reviews. The authors identify employee-discussed EX-factors and quantify the associated sentiments and importance.Design/methodology/approachThis paper employs Latent Dirichlet Allocation on the online employee reviews to identify the key EX-factors. The authors probe sentiments and importance associated with key EX-factors using sentiment analysis, importance analysis, regression analysis and dominance analysis.FindingsThe result of topic modeling identifies 20 EX-factors that shape overall EX. While skill development plays a major role in shaping overall EX, employees perceived Salary and Growth as the most important EX-factor and expressed negative sentiments during the pandemic. Employee sentiments significantly influence overall EX.Practical implicationsWhen employees have extensive change experience, managers should consider various facets of EX to manage the smooth change and deliver a better EX. This research offers key EX-factors to be considered by managers while dealing with employees. Online employee reviews websites are recommended to include the identified key EX-factors to comprehend the holistic EX.Originality/valueThis study contributes to the growing literature on the employee experience as a concept by identifying various EX-factors. The authors expand the extant EX scales by identifying an inclusive and updated set of EX-factors.

show abstract

Section: Topic Modellingmentioning

confidence: 99%

Decoding employee experiences during pandemic through online employee reviews: insights to organizations

Joshi

Sekar

Das

2023

View full text Add to dashboard Cite

show abstract

“…LDA topic analysis requires input parameters to increase their performance, for example, the number of topics ( k ) and a prior Dirichlet topic distribution in the function. However, finding the best number of topics is challenging ( 43 ), especially if there is no prior knowledge about the data. The LDA topic model has a problem in that it does not give the optimal number of topics for the text itself, the exact number of topics being determined by the model user in other ways ( 44 ).…”

Section: Methodsmentioning

confidence: 99%

Impacts of Precautionary and Opportunistic Buying Behaviors and Supply Issues on Supply Chain Resilience During the COVID-19 Pandemic

Rivera-Royero

Jaller

Jenn

2022

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

This paper proposes a text analytics approach using dictionary-based clustering, word counting data, and discrete regression modeling to study the relationship between demand behaviors and supply issues affecting supply chain resilience during the first stages of the COVID-19 pandemic. The work used news and media articles gathered from the LexisNexis database covering 5 months between February and July 2020. The data analyses describe the general patterns in the news texts by using text mining techniques, and the methodology describes the relationship between consumer behavior, supply chain issues, and the reduced level of service shown during the study period. Demand behaviors include precautionary and opportunistic buying, which affected many countries and could be the result of a lack of perceived control and other factors; for example, near-empty shelves of certain products could have prompted consumers to increasingly look for comparable products, driving up demand. Additionally, the method explored the potential effect of strategies to mitigate impacts on the resilience of supply chains. The results confirmed that buying behaviors and a reduction in the capacity of the supply chain led to a lower level of service being perceived by consumers, however, resilience strategies were found to mitigate the impact of such capacity reductions. Empirical analyses showed that the proposed approach, using data extracted from the news, could identify and represent impacts consistent with expectations from the supply chain field under disruptions, and quantify the magnitude of the impacts as the pandemic evolved, providing more information for decision making.

show abstract

“…Another possible approach would have consisted in performing topic modeling, for instance by ways of a Latent Dirichlet Allocation on the word frequency matrix, to then infer a distribution of topics for every county. It is, however, more computationally intensive, and poses questions about the selection of the number of topics, their interpretability, and their internal coherence (Arun et al, 2010;Hasan et al, 2021). In a case like ours where documents are so large (aggregating all tweets in a county), it is far from obvious to select a number of topics such that there is little overlap between them and to know that these topics are actually representative of the dataset as a whole.…”

Section: Datasetmentioning

confidence: 99%

American cultural regions mapped through the lexical analysis of social media

Louf

Gonçalves

Ramasco

et al. 2023

Humanit Soc Sci Commun

View full text Add to dashboard Cite

Cultural areas represent a useful concept that cross-fertilizes diverse fields in social sciences. Knowledge of how humans organize and relate their ideas and behavior within a society can help us to understand our actions and attitudes toward different issues. However, the selection of common traits that shape a cultural area is somewhat arbitrary. What is needed is a method that can leverage the massive amounts of data coming online, especially through social media, to identify cultural regions without ad-hoc assumptions, biases, or prejudices. This work takes a crucial step in this direction by introducing a method to infer cultural regions based on the automatic analysis of large datasets from microblogging posts. The approach presented here is based on the principle that cultural affiliation can be inferred from the topics that people discuss among themselves. Specifically, regional variations in written discourse are measured in American social media. From the frequency distributions of content words in geotagged tweets, the regional hotspots of words’ usage are found, and from there, principal components of regional variation are derived. Through a hierarchical clustering of the data in this lower-dimensional space, this method yields clear cultural areas and the topics of discussion that define them. It uncovers a manifest North–South separation, which is primarily influenced by the African American culture, and further contiguous (East–West) and non-contiguous divisions that provide a comprehensive picture of modern American cultural areas.

show abstract

Normalized Approach to Find Optimal Number of Topics in Latent Dirichlet Allocation (LDA)

Cited by 49 publications

References 15 publications

Decoding employee experiences during pandemic through online employee reviews: insights to organizations

Decoding employee experiences during pandemic through online employee reviews: insights to organizations

Impacts of Precautionary and Opportunistic Buying Behaviors and Supply Issues on Supply Chain Resilience During the COVID-19 Pandemic

American cultural regions mapped through the lexical analysis of social media

Contact Info

Product

Resources

About