164 lines
9.4 KiB
Markdown
164 lines
9.4 KiB
Markdown
file-path:: file://C:\Users\david\Zotero/storage/59BGRZB6/Turki et al. - Recommending scholarly articles to monitor COVID-1.pdf
|
|
file:: [Turki et al. - Recommending scholarly articles to monitor COVID-1.pdf](file://C:\Users\david\Zotero/storage/59BGRZB6/Turki et al. - Recommending scholarly articles to monitor COVID-1.pdf)
|
|
title:: hls__Turki et al. - Recommending scholarly articles to monitor COVID-1
|
|
|
|
- generate precise topic clusters reflecting the social trends about the pandemic at a low cost
|
|
ls-type:: annotation
|
|
hl-page:: 1
|
|
id:: 6305cf53-e448-46d8-9532-2ad8c8e3c5c6
|
|
- the analysis of social media interactions related to the disease can be very useful to assess the general perception of the concerned disease by a local population, identify rumors and conspiracy theories about the COVID-19 pandemic, and track the spread and effect of official information, news and guidelines about the disease outbreak among a specific community
|
|
ls-type:: annotation
|
|
hl-page:: 1
|
|
id:: 6305cff4-58ab-4f47-9612-7cf834d839c0
|
|
- problem
|
|
ls-type:: annotation
|
|
hl-page:: 1
|
|
id:: 6305d335-4222-4cc2-8a08-d58a9d54dc48
|
|
- d facilitating rapid datadriven decision-making to face any encountered societa
|
|
ls-type:: annotation
|
|
hl-page:: 1
|
|
id:: 6305d341-450e-4a2a-915a-e1851041384d
|
|
- black boxes that cannot be easily debugged and adjusted transparentl
|
|
ls-type:: annotation
|
|
hl-page:: 1
|
|
id:: 6305d360-4dd9-4ede-a2d7-63eb7c8ff0cf
|
|
- trustworthy artificial intelligence applications allowing certain and responsible decision-making related to COVID-19
|
|
ls-type:: annotation
|
|
hl-page:: 1
|
|
id:: 6305d36b-fd13-4a06-a369-fa73c0bc8403
|
|
- all the developing nations do not have the means to purchase expensive GPUs and cloud services to allow complex computations for better accuracy of machine learning algorithms
|
|
ls-type:: annotation
|
|
hl-page:: 1
|
|
id:: 6305d38a-eb02-4493-aabc-cf3fc5e2766a
|
|
- many countries in the Global South have limited human development (HDI < 0.7) and incomes (Nominal GDP per capita < 2000 USD
|
|
ls-type:: annotation
|
|
hl-page:: 1
|
|
id:: 6305d395-86ef-46f1-8339-217410c8196d
|
|
- African countries published less than 2600 papers in 2020, scientifically developed countries like China and the United States of America published more than 10000 papers
|
|
ls-type:: annotation
|
|
hl-page:: 1
|
|
id:: 6305d3ed-5202-414d-97f1-340f60e54f49
|
|
- topic modeling of the topics of interest related to the ongoing disease outbreak for a local population at a low cost
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305d418-5e42-4a96-b30f-8c1063eaf099
|
|
- we propose to use the returned topic clusters to recommend scholarly publications that can be used by health professionals and authorities to fight widespread misinformation and provide interesting accurate guidelines for their communities concerned by COVID-19 through the data mining of PubMed Central2, a database of open access biomedical research publications available online
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305d9bd-078e-437c-9f9a-46e021bd81fd
|
|
- precise topic modeling of the social network interactions related to the COVID-19
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305d9f7-ee67-48b9-b29d-715ae6efa72c
|
|
- generation of useful scholarly publications for tracking and adjusting social thoughts about the disease outbreak
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305d9fe-579d-463c-97e4-3a75700ab512
|
|
- Social Network Analysis for Crisis Management
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305da89-3e00-4430-a18b-162b01ef5641
|
|
- The ability of these online platforms (e.g., Facebook) to establish virtual connections between individuals has permitted these websites to have billions of users within a few years of work [ 13 ]. Nowadays, thanks to their growth, social networks provide real-time big data about various aspects of human concerns including political and health crises. This resulted in the emergence of a significant research trend of using social network interactions to track social responses to crises.
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305dae5-07ce-484d-b114-0bcf157e9ba2
|
|
- users in social networks tend to react to an ongoing crisis by seeking, posting, sharing, and interacting with information about the phenomenon in a massive way (so-called infodemic)
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305db83-55fc-464f-b75b-0c6f41a8a6e1
|
|
- As a result, the application of computer methods to analyze social network data can reliably reflect the sentiments and thoughts of a given community about the considered crisis and help identify and predict the geographical and socio-economic evolution of the phenomenon
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305dc57-492a-4277-8804-7cbcc64e16fd
|
|
- Social network analysis can be a valuable tool for detecting and eliminating inconsistent posts spreading misinformation and rumors across social networking sites leaving room for accurate posts and knowledge about the considered topic to get more disseminated
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305dc69-e0cd-4b66-a5e8-88d5c7501aac
|
|
- combination of both data types coupled with social network analysis enables the development of knowledge-based systems to identify the main trendy topics for users as well as to measure the similarity between scholarly publications and user interests [ 26 ]. The outcomes of such intelligent systems will allow the generation of accurate recommendations of scholarly articles to meet the needs of assessed users
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305dd07-642a-40b0-99d6-e41f81a6ff96
|
|
- generated user interest profile
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305ddd8-f38a-4c16-8b1f-6acdb2955cb6
|
|
- semantic similarity measures
|
|
ls-type:: annotation
|
|
hl-page:: 2
|
|
id:: 6305dde0-4a9f-456c-a6be-2e5c77ed29bc
|
|
- here are some collaborative filtering approaches for the recommendation of scholarly publications for a given user based on the readings and behaviors of other users
|
|
ls-type:: annotation
|
|
hl-page:: 3
|
|
id:: 6305de59-82ac-44dd-8617-0c0a2323c410
|
|
- Despite the variety of scholarly publication recommender systems, quite all of them propose further readings based on the interests of a particular user and not of a global social community.
|
|
ls-type:: annotation
|
|
hl-page:: 3
|
|
id:: 6305de7b-2dc0-40a9-8b58-439c0a5f5931
|
|
- Despite the variety of scholarly publication recommender systems, quite all of them propose further readings based on the interests of a particular user and not of a global social community.
|
|
ls-type:: annotation
|
|
hl-page:: 3
|
|
id:: 6305de95-1fd9-4633-9c3a-dd344d00298d
|
|
- or single users of social networks
|
|
ls-type:: annotation
|
|
hl-page:: 3
|
|
id:: 6305def3-402f-4e05-adc3-c9ca211aa62d
|
|
- cannot be efficient in the situation of a broad crisis like COVID-19 when specialized information is requested on a large scale
|
|
ls-type:: annotation
|
|
hl-page:: 3
|
|
id:: 6305df01-afdd-4118-9623-75ed9eb91f88
|
|
- Global South countries, especially Tunisia
|
|
ls-type:: annotation
|
|
hl-page:: 3
|
|
id:: 6305df3c-0a55-44b7-a7c2-d367ff63718d
|
|
- feature selection from the input data to generate their outputs for data analysis, validation, and recommendation
|
|
ls-type:: annotation
|
|
hl-page:: 3
|
|
id:: 6305df54-ff00-4030-bfb7-60dc69468bbc
|
|
- feature selection is mostly probabilistic and depends on the statistical features of the input data
|
|
ls-type:: annotation
|
|
hl-page:: 3
|
|
id:: 6305df5f-a5e4-4f6d-bdf9-381a96964109
|
|
- Author concept is used for presenting users across online social networks and their related metadata such as name, age, interest, etc.
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
id:: 6305e0df-6c94-4827-b5c6-d03c4d36d1bb
|
|
- this heterogeneity requires specific treatment for each social network for identifying COVID-19-related social entities.
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
id:: 6305e133-a2df-4a47-9b9a-0f94b6359003
|
|
- Figure 1: Architecture of the recommender system for scholar publications to monitor the pandemic COVD-19 based on social data analysis).
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
id:: 6305e20e-14a7-40eb-8dda-4da889cd84be
|
|
- Best Match
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
id:: 6305e8a2-8b37-4ade-afe7-2ec68945884c
|
|
- , a high-performance pre-trained model that classifies publications using a variety of characteristics extracted from queries and documents
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
id:: 6305e8af-912a-41e4-a162-3ba231a0a58c
|
|
- agreement between the ten first results of a query with the ones of the baseline
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
id:: 6305e8fa-cb20-43ec-964a-806e19ed69a3
|
|
- hundred first results of a query with the ones of the baseline
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
id:: 6305e902-80e8-4d86-bbf0-394f797da48d
|
|
- it is revealed in the two tables that the query runtime tends to be largely shortened when the most specific keyword is put first in the query
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
id:: 6305e93b-c7dd-4da7-8b39-18c495b56152
|
|
- This demonstrates that keyword duplication can be practically used to emphasize one keyword in the query over other ones
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
id:: 6305e97d-f0f0-4ec0-86fd-54c7117c2dcf
|
|
- topic models have emerged as statistical models allowing to reveal the abstract clusters of terms that represent a collection of documents
|
|
ls-type:: annotation
|
|
hl-page:: 6
|
|
id:: 6305e991-602c-430b-a613-bb67a257daa5
|
|
- This concurrent research is considered as a recommender system to provide scholarly publications to monitor and track the COVID-19 pandemic which depends on the analysis of the data of the social platforms Facebook and Twitter.
|
|
ls-type:: annotation
|
|
hl-page:: 7
|
|
id:: 6305e9c0-c6d3-4770-b5d3-e49960d72a32 |