2020
DOI: 10.1016/j.visinf.2020.04.005
|View full text |Cite
|
Sign up to set email alerts
|

A visual analytics system for multi-model comparison on clinical data predictions

Abstract: There is a growing trend of applying machine learning methods to medical datasets in order to predict patients' future status. Although some of these methods achieve high performance, challenges still exist in comparing and evaluating different models through their interpretable information. Such analytics can help clinicians improve evidence-based medical decision making. In this work, we develop a visual analytics system that compares multiple models' prediction criteria and evaluates their consistency. With… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
16
0

Year Published

2020
2020
2023
2023

Publication Types

Select...
5
2
1

Relationship

0
8

Authors

Journals

citations
Cited by 27 publications
(16 citation statements)
references
References 45 publications
0
16
0
Order By: Relevance
“…Zintgraf, Cohen, Adel, and Welling (2017) present a method that visualizes which pixels of an input image are evidence for or evidence against a node in a deep neural network. Moreover, Li, Fujiwara, Choi, Kim, and Ma (2020) focus on visualizing features to explain the reasoning of a model. The approach is used to explain and compare different models for clinical data beyond their accuracy scores.…”
Section: Box 1 Visual Analytics and Interpretabilitymentioning
confidence: 99%
“…Zintgraf, Cohen, Adel, and Welling (2017) present a method that visualizes which pixels of an input image are evidence for or evidence against a node in a deep neural network. Moreover, Li, Fujiwara, Choi, Kim, and Ma (2020) focus on visualizing features to explain the reasoning of a model. The approach is used to explain and compare different models for clinical data beyond their accuracy scores.…”
Section: Box 1 Visual Analytics and Interpretabilitymentioning
confidence: 99%
“…Another example from pain research is how subsymbolic AI can be subjected to further analysis to extract the individual decision process for each case, or how it can be complemented by symbolic AI that provides understandable explanations for group assignment, although the exact decision process may differ from that of subsymbolic AI [37]. This has been elaborated in more detail in a visual analysis system for multimodel comparison of predictions for clinical data [75]. The system allows comparison and evaluation of different AI models based on their interpretable information, with the goal of assisting clinicians in decision making.…”
Section: Accessibilitymentioning
confidence: 99%
“…G3 Comparative analysis of RF models. Another design goal is the ability to have a comparative analysis between two or more RF models to assist model developers in selecting reliable models [50]. An RF model is built using various parameters such as the number of trees, splitting criteria, maximum depth of a tree, and the maximum number of features to be considered during a split.…”
Section: G2 Local Interpretation By Preserving the Global Context Loc...mentioning
confidence: 99%