2010
DOI: 10.1007/978-3-642-13022-9_31
|View full text |Cite
|
Sign up to set email alerts
|

An Environment for Data Analysis in Biomedical Domain: Information Extraction for Decision Support Systems

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
3
0

Year Published

2020
2020
2023
2023

Publication Types

Select...
4
1

Relationship

0
5

Authors

Journals

citations
Cited by 5 publications
(5 citation statements)
references
References 14 publications
0
3
0
Order By: Relevance
“…The revocation of the toxic messages stands out, with 95% of the results being retrieved. The F-score, a harmonic weighted average of the precision and revocation scores using a constant β to determine the weight given to each score, can also be used to measure an information retrieval model's performance [18]. The β chosen for this experiment's F-score was of one (1), such that revocation and precision were given an equal weight [18], resulting in a relatively high F-score for both the toxic and non-toxic messages, receiving 0.84 and 0.89, respectively.…”
Section: Discussionmentioning
confidence: 99%
See 1 more Smart Citation
“…The revocation of the toxic messages stands out, with 95% of the results being retrieved. The F-score, a harmonic weighted average of the precision and revocation scores using a constant β to determine the weight given to each score, can also be used to measure an information retrieval model's performance [18]. The β chosen for this experiment's F-score was of one (1), such that revocation and precision were given an equal weight [18], resulting in a relatively high F-score for both the toxic and non-toxic messages, receiving 0.84 and 0.89, respectively.…”
Section: Discussionmentioning
confidence: 99%
“…The F-score, a harmonic weighted average of the precision and revocation scores using a constant β to determine the weight given to each score, can also be used to measure an information retrieval model's performance [18]. The β chosen for this experiment's F-score was of one (1), such that revocation and precision were given an equal weight [18], resulting in a relatively high F-score for both the toxic and non-toxic messages, receiving 0.84 and 0.89, respectively. A possible explanation for the model's success in classifying toxic messages is the repeated use of tokens that were always attributed as toxic, including racist and homophobic slurs, along with toxic game-related slang like the term "ez".…”
Section: Discussionmentioning
confidence: 99%
“…Os experimentos foram realizados em duas etapas: a primeira com as bases completas e a segunda com um subconjunto de variáveis filtradas a partir da aplicac ¸ão da técnica de selec ¸ão de atributos. A métrica de avaliac ¸ão utilizada foi a acurácia, esta indica a porcentagem de acerto de um classificador, ou seja, quanto maior a acurácia, melhor é o modelo gerado [Matos et al 2009].…”
Section: Modelagemunclassified
“…Foi realizado o treinamento do classificador, resultando na geração de uma matriz de confusão, que proporciona uma avaliação eficaz do modelo de classificação, ao mostrar o número de classificações corretas e as classificações preditas para cada classe em um determinado conjunto de exemplos (MATOS et al, 2009). Essas classificações são: verdadeiros positivos (VP), onde o modelo classifica corretamente as instâncias como pertencentes à classe em questão; verdadeiros negativos (VN), nos quais o modelo classificar corretamente as instâncias como não pertencentes à classe; falsos positivos (FP), quando o modelo incorretamente classifica uma instância como pertencente a uma classe específica, quando na verdade não pertence; e falsos negativos (FN), onde uma instância é equivocadamente classificada como não pertencente a uma determinada classe, embora ela realmente pertença.…”
Section: Metodologiaunclassified