EFI: A Toolbox for Feature Importance Fusion and Interpretation in Python

Kumar, Aayush; Mase, Jimiama Mafeni; Rengasamy, Divish; Rothwell, Benjamin; Torres, Mercedes Torres; Winkler, David A.; Figueredo, Grazziela P.

doi:10.1007/978-3-031-25891-6_19

Cited by 2 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to evaluate the possibility of determining emotional state from the data of specific groups of sensors for decision-tree-based machine learning models, the feature_importance method implemented in Python [ 58 ] was applied. The value of feature_importance is determined as follows:

where c 1 and c 2 are the total weight of the objects in the left and right leaves, respectively; and v 1 and v 2 are the value of Formula (1) for the left and right leaves at the previous tree level.…”

Section: Processing Of Measurement Resultsmentioning

confidence: 99%

“…In order to evaluate the possibility of determining emotional state from the data of specific groups of sensors for decision-tree-based machine learning models, the fea-ture_importance method implemented in Python [58] was applied. The value of fea-ture_importance is determined as follows: This parameter allows us to identify the sensors, the triggering of which was most active in certain emotions.…”

Section: Processing Of Measurement Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Assessment of a Person’s Emotional State Based on His or Her Posture Parameters

Shichkina,

Bureneva,

Salaurov

et al. 2023

Sensors

View full text Add to dashboard Cite

This article is devoted to the study of the correlation between the emotional state of a person and the posture of his or her body in the sitting position. In order to carry out the study, we developed the first version of the hardware-software system based on a posturometric armchair, allowing the characteristics of the posture of a sitting person to be evaluated using strain gauges. Using this system, we revealed the correlation between sensor readings and human emotional states. We showed that certain readings of a sensor group are formed for a certain emotional state of a person. We also found that the groups of triggered sensors, their composition, their number, and their location are related to the states of a particular person, which led to the need to build personalized digital pose models for each person. The intellectual component of our hardware–software complex is based on the concept of co-evolutionary hybrid intelligence. The system can be used during medical diagnostic procedures and rehabilitation processes, as well as in controlling people whose professional activity is connected with increased psycho-emotional load and can cause cognitive disorders, fatigue, and professional burnout and can lead to the development of diseases.

show abstract