A deficiency of prescriptive analytics—No perfect predicted value or predicted distribution exists
Researchers and industrial practitioners are now interested in combining machine learning (ML) and operations research and management science to develop prescriptive analytics frameworks. By and large, a single value or a discrete distribution with a finite number of scenarios is predicted using an...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIMS Press
2022-07-01
|
| Series: | Electronic Research Archive |
| Subjects: | |
| Online Access: | https://www.aimspress.com/article/doi/10.3934/era.2022183?viewType=HTML |
| _version_ | 1811233535555534848 |
|---|---|
| author | Shuaian Wang Xuecheng Tian Ran Yan Yannick Liu |
| author_facet | Shuaian Wang Xuecheng Tian Ran Yan Yannick Liu |
| author_sort | Shuaian Wang |
| collection | DOAJ |
| description | Researchers and industrial practitioners are now interested in combining machine learning (ML) and operations research and management science to develop prescriptive analytics frameworks. By and large, a single value or a discrete distribution with a finite number of scenarios is predicted using an ML model with an unknown parameter; the value or distribution is then fed into an optimization model with the unknown parameter to prescribe an optimal decision. In this paper, we prove a deficiency of prescriptive analytics, i.e., that no perfect predicted value or perfect predicted distribution exists in some cases. To illustrate this phenomenon, we consider three different frameworks of prescriptive analytics, namely, the predict-then-optimize framework, smart predict-then-optimize framework and weighted sample average approximation (w-SAA) framework. For these three frameworks, we use examples to show that prescriptive analytics may not be able to prescribe a full-information optimal decision, i.e., the optimal decision under the assumption that the distribution of the unknown parameter is given. Based on this finding, for practical prescriptive analytics problems, we suggest comparing the prescribed results among different frameworks to determine the most appropriate one. |
| first_indexed | 2024-04-12T11:21:36Z |
| format | Article |
| id | doaj.art-e58c8bd6bf7c4bb783b0371b893286ba |
| institution | Directory Open Access Journal |
| issn | 2688-1594 |
| language | English |
| last_indexed | 2024-04-12T11:21:36Z |
| publishDate | 2022-07-01 |
| publisher | AIMS Press |
| record_format | Article |
| series | Electronic Research Archive |
| spelling | doaj.art-e58c8bd6bf7c4bb783b0371b893286ba2022-12-22T03:35:22ZengAIMS PressElectronic Research Archive2688-15942022-07-0130103586359410.3934/era.2022183A deficiency of prescriptive analytics—No perfect predicted value or predicted distribution existsShuaian Wang 0Xuecheng Tian1Ran Yan2Yannick Liu31. Department of Logistics and Maritime Studies, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong1. Department of Logistics and Maritime Studies, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong1. Department of Logistics and Maritime Studies, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong2. Faculty of Business, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong KongResearchers and industrial practitioners are now interested in combining machine learning (ML) and operations research and management science to develop prescriptive analytics frameworks. By and large, a single value or a discrete distribution with a finite number of scenarios is predicted using an ML model with an unknown parameter; the value or distribution is then fed into an optimization model with the unknown parameter to prescribe an optimal decision. In this paper, we prove a deficiency of prescriptive analytics, i.e., that no perfect predicted value or perfect predicted distribution exists in some cases. To illustrate this phenomenon, we consider three different frameworks of prescriptive analytics, namely, the predict-then-optimize framework, smart predict-then-optimize framework and weighted sample average approximation (w-SAA) framework. For these three frameworks, we use examples to show that prescriptive analytics may not be able to prescribe a full-information optimal decision, i.e., the optimal decision under the assumption that the distribution of the unknown parameter is given. Based on this finding, for practical prescriptive analytics problems, we suggest comparing the prescribed results among different frameworks to determine the most appropriate one.https://www.aimspress.com/article/doi/10.3934/era.2022183?viewType=HTMLprescriptive analyticspredict-then-optimizesmart predict-then-optimizeweighted sample average approximation |
| spellingShingle | Shuaian Wang Xuecheng Tian Ran Yan Yannick Liu A deficiency of prescriptive analytics—No perfect predicted value or predicted distribution exists Electronic Research Archive prescriptive analytics predict-then-optimize smart predict-then-optimize weighted sample average approximation |
| title | A deficiency of prescriptive analytics—No perfect predicted value or predicted distribution exists |
| title_full | A deficiency of prescriptive analytics—No perfect predicted value or predicted distribution exists |
| title_fullStr | A deficiency of prescriptive analytics—No perfect predicted value or predicted distribution exists |
| title_full_unstemmed | A deficiency of prescriptive analytics—No perfect predicted value or predicted distribution exists |
| title_short | A deficiency of prescriptive analytics—No perfect predicted value or predicted distribution exists |
| title_sort | deficiency of prescriptive analytics no perfect predicted value or predicted distribution exists |
| topic | prescriptive analytics predict-then-optimize smart predict-then-optimize weighted sample average approximation |
| url | https://www.aimspress.com/article/doi/10.3934/era.2022183?viewType=HTML |
| work_keys_str_mv | AT shuaianwang adeficiencyofprescriptiveanalyticsnoperfectpredictedvalueorpredicteddistributionexists AT xuechengtian adeficiencyofprescriptiveanalyticsnoperfectpredictedvalueorpredicteddistributionexists AT ranyan adeficiencyofprescriptiveanalyticsnoperfectpredictedvalueorpredicteddistributionexists AT yannickliu adeficiencyofprescriptiveanalyticsnoperfectpredictedvalueorpredicteddistributionexists AT shuaianwang deficiencyofprescriptiveanalyticsnoperfectpredictedvalueorpredicteddistributionexists AT xuechengtian deficiencyofprescriptiveanalyticsnoperfectpredictedvalueorpredicteddistributionexists AT ranyan deficiencyofprescriptiveanalyticsnoperfectpredictedvalueorpredicteddistributionexists AT yannickliu deficiencyofprescriptiveanalyticsnoperfectpredictedvalueorpredicteddistributionexists |