SCIENCE@home

Prospect Value Function Construction for Pattern Recognition Algorithm’s Accuracy Metric

Jun He,Ruigang Fu,Guoyan Wang

Published 2022 in ACM Cloud and Autonomic Computing Conference

ABSTRACT

Accuracy metric of pattern recognition algorithm has a small amount of randomness because of train/test data selection and their sample sizes limitation. We proposed a prospect value function constructing method for pattern recognition algorithm by translating its accuracy to prospect value in order to avoid the potential random of accuracy. The normalized prospect value could also be used for further multi-attribute decision making. We analyzed some important steps in constructing the prospect value function of a pattern recognition algorithm, such as reference point selection, value function development, decision weight calculation and so on. The numerical application shows that the prospect value function constructed in this paper has a diminishing marginal. Further discussions conform that our method is fit for most pattern recognition algorithms’ accuracy domain and insensitive to test data sample size variation, i.e., the pattern recognition algorithms’ accuracy improvement can be measured quantitatively by its prospect value function.

PUBLICATION RECORD

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-20 of 20 references · Page 1 of 1

CITED BY