Regional analysis of socioeconomic indicators and federated learning performance gain.

<p>The table presents a systematic comparison across Brazil’s five macro-regions, linking key socioeconomic and health indicators with the study’s sample characteristics (number of hospitals and average patient cohort size per hospital). The primary outcome, the mean performance gain from fede...

Mô tả đầy đủ

Đã lưu trong:
Chi tiết về thư mục
Tác giả chính: Roberta Moreira Wichmann (14259316) (author)
Tác giả khác: Murilo Afonso Robiati Bigoto (22676715) (author), Alexandre Dias Porto Chiavegatto Filho (14259328) (author)
Được phát hành: 2025
Những chủ đề:
Các nhãn: Thêm thẻ
Không có thẻ, Là người đầu tiên thẻ bản ghi này!
_version_ 1849927641224708096
author Roberta Moreira Wichmann (14259316)
author2 Murilo Afonso Robiati Bigoto (22676715)
Alexandre Dias Porto Chiavegatto Filho (14259328)
author2_role author
author
author_facet Roberta Moreira Wichmann (14259316)
Murilo Afonso Robiati Bigoto (22676715)
Alexandre Dias Porto Chiavegatto Filho (14259328)
author_role author
dc.creator.none.fl_str_mv Roberta Moreira Wichmann (14259316)
Murilo Afonso Robiati Bigoto (22676715)
Alexandre Dias Porto Chiavegatto Filho (14259328)
dc.date.none.fl_str_mv 2025-11-24T18:35:39Z
dc.identifier.none.fl_str_mv 10.1371/journal.pcbi.1013695.s005
dc.relation.none.fl_str_mv https://figshare.com/articles/dataset/Regional_analysis_of_socioeconomic_indicators_and_federated_learning_performance_gain_/30697800
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Cell Biology
Ecology
Cancer
Science Policy
Biological Sciences not elsewhere classified
Chemical Sciences not elsewhere classified
Information Systems not elsewhere classified
via parameter aggregation
using ensemble aggregation
strong inverse correlation
maximum performance degradation
benefits requires caution
beneficial collaborative effect
>&# 916 ;</
19 mortality using
19 mortality prediction
local patient volume
statistically certain benefits
statistical significance showed
0273 ]) due
auc crossed zero
study validates fl
fl models demonstrated
performance gain (<
evaluated federated learning
div >< p
collaborative model ’
auc minus auc
local data scarcity
ci [– 0
gain ’
federated learning
local validation
universally certain
statistical advantage
6307 ])
auc across
– 0
volatile algorithm
smallest hospital
particularly evident
network (<
multicentric sample
multicenter sample
limited institutions
layer perceptron
iid ).
high sensitivity
findings underscore
every site
enabling mechanism
despite achieving
confidence intervals
500 ).
022 patients
dc.title.none.fl_str_mv Regional analysis of socioeconomic indicators and federated learning performance gain.
dc.type.none.fl_str_mv Dataset
info:eu-repo/semantics/publishedVersion
dataset
description <p>The table presents a systematic comparison across Brazil’s five macro-regions, linking key socioeconomic and health indicators with the study’s sample characteristics (number of hospitals and average patient cohort size per hospital). The primary outcome, the mean performance gain from federated learning (<i>Δ</i>AUC = AUC - AUC), is presented for each of the three models: Logistic Regression (LR), Multilayer Perceptron (MLP), and Random Forest (RF). The mean <i>Δ</i>AUC values for each region were calculated by averaging the hospital-specific mean <i>Δ</i>AUCs reported in S1 Table (for Logistic Regression), <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1013695#pcbi.1013695.s003" target="_blank">S2 Table</a> (for MLP), and <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1013695#pcbi.1013695.s004" target="_blank">S3 Table</a> (for Random Forest) across all hospitals within that geographical region. This allows for a higher-level investigation of the relationship between regional characteristics and the benefits of the federated approach.</p> <p>(XLSX)</p>
eu_rights_str_mv openAccess
id Manara_07cd2407db769c1a6b9df1e79d18abe7
identifier_str_mv 10.1371/journal.pcbi.1013695.s005
network_acronym_str Manara
network_name_str ManaraRepo
oai_identifier_str oai:figshare.com:article/30697800
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Regional analysis of socioeconomic indicators and federated learning performance gain.Roberta Moreira Wichmann (14259316)Murilo Afonso Robiati Bigoto (22676715)Alexandre Dias Porto Chiavegatto Filho (14259328)Cell BiologyEcologyCancerScience PolicyBiological Sciences not elsewhere classifiedChemical Sciences not elsewhere classifiedInformation Systems not elsewhere classifiedvia parameter aggregationusing ensemble aggregationstrong inverse correlationmaximum performance degradationbenefits requires cautionbeneficial collaborative effect>&# 916 ;</19 mortality using19 mortality predictionlocal patient volumestatistically certain benefitsstatistical significance showed0273 ]) dueauc crossed zerostudy validates flfl models demonstratedperformance gain (<evaluated federated learningdiv >< pcollaborative model ’auc minus auclocal data scarcityci [– 0gain ’federated learninglocal validationuniversally certainstatistical advantage6307 ])auc across– 0volatile algorithmsmallest hospitalparticularly evidentnetwork (<multicentric samplemulticenter samplelimited institutionslayer perceptroniid ).high sensitivityfindings underscoreevery siteenabling mechanismdespite achievingconfidence intervals500 ).022 patients<p>The table presents a systematic comparison across Brazil’s five macro-regions, linking key socioeconomic and health indicators with the study’s sample characteristics (number of hospitals and average patient cohort size per hospital). The primary outcome, the mean performance gain from federated learning (<i>Δ</i>AUC = AUC - AUC), is presented for each of the three models: Logistic Regression (LR), Multilayer Perceptron (MLP), and Random Forest (RF). The mean <i>Δ</i>AUC values for each region were calculated by averaging the hospital-specific mean <i>Δ</i>AUCs reported in S1 Table (for Logistic Regression), <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1013695#pcbi.1013695.s003" target="_blank">S2 Table</a> (for MLP), and <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1013695#pcbi.1013695.s004" target="_blank">S3 Table</a> (for Random Forest) across all hospitals within that geographical region. This allows for a higher-level investigation of the relationship between regional characteristics and the benefits of the federated approach.</p> <p>(XLSX)</p>2025-11-24T18:35:39ZDatasetinfo:eu-repo/semantics/publishedVersiondataset10.1371/journal.pcbi.1013695.s005https://figshare.com/articles/dataset/Regional_analysis_of_socioeconomic_indicators_and_federated_learning_performance_gain_/30697800CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/306978002025-11-24T18:35:39Z
spellingShingle Regional analysis of socioeconomic indicators and federated learning performance gain.
Roberta Moreira Wichmann (14259316)
Cell Biology
Ecology
Cancer
Science Policy
Biological Sciences not elsewhere classified
Chemical Sciences not elsewhere classified
Information Systems not elsewhere classified
via parameter aggregation
using ensemble aggregation
strong inverse correlation
maximum performance degradation
benefits requires caution
beneficial collaborative effect
>&# 916 ;</
19 mortality using
19 mortality prediction
local patient volume
statistically certain benefits
statistical significance showed
0273 ]) due
auc crossed zero
study validates fl
fl models demonstrated
performance gain (<
evaluated federated learning
div >< p
collaborative model ’
auc minus auc
local data scarcity
ci [– 0
gain ’
federated learning
local validation
universally certain
statistical advantage
6307 ])
auc across
– 0
volatile algorithm
smallest hospital
particularly evident
network (<
multicentric sample
multicenter sample
limited institutions
layer perceptron
iid ).
high sensitivity
findings underscore
every site
enabling mechanism
despite achieving
confidence intervals
500 ).
022 patients
status_str publishedVersion
title Regional analysis of socioeconomic indicators and federated learning performance gain.
title_full Regional analysis of socioeconomic indicators and federated learning performance gain.
title_fullStr Regional analysis of socioeconomic indicators and federated learning performance gain.
title_full_unstemmed Regional analysis of socioeconomic indicators and federated learning performance gain.
title_short Regional analysis of socioeconomic indicators and federated learning performance gain.
title_sort Regional analysis of socioeconomic indicators and federated learning performance gain.
topic Cell Biology
Ecology
Cancer
Science Policy
Biological Sciences not elsewhere classified
Chemical Sciences not elsewhere classified
Information Systems not elsewhere classified
via parameter aggregation
using ensemble aggregation
strong inverse correlation
maximum performance degradation
benefits requires caution
beneficial collaborative effect
>&# 916 ;</
19 mortality using
19 mortality prediction
local patient volume
statistically certain benefits
statistical significance showed
0273 ]) due
auc crossed zero
study validates fl
fl models demonstrated
performance gain (<
evaluated federated learning
div >< p
collaborative model ’
auc minus auc
local data scarcity
ci [– 0
gain ’
federated learning
local validation
universally certain
statistical advantage
6307 ])
auc across
– 0
volatile algorithm
smallest hospital
particularly evident
network (<
multicentric sample
multicenter sample
limited institutions
layer perceptron
iid ).
high sensitivity
findings underscore
every site
enabling mechanism
despite achieving
confidence intervals
500 ).
022 patients