CN113506021A - Index dimensionless processing method for comprehensive evaluation - Google Patents
Index dimensionless processing method for comprehensive evaluation Download PDFInfo
- Publication number
- CN113506021A CN113506021A CN202110845817.9A CN202110845817A CN113506021A CN 113506021 A CN113506021 A CN 113506021A CN 202110845817 A CN202110845817 A CN 202110845817A CN 113506021 A CN113506021 A CN 113506021A
- Authority
- CN
- China
- Prior art keywords
- evaluation
- index
- evaluation data
- data
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000011156 evaluation Methods 0.000 title claims abstract description 103
- 238000003672 processing method Methods 0.000 title claims abstract description 15
- 230000002159 abnormal effect Effects 0.000 claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 14
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06311—Scheduling, planning or task assignment for a person or group
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Mining
Abstract
The invention discloses an index dimensionless processing method for comprehensive evaluation, which relates to the technical field of comprehensive evaluation and comprises the following steps: constructing original evaluation data sequences of different models to be evaluated by evaluation indexes; segmenting the original evaluation data sequence corresponding to each evaluation index according to the abnormal value and the non-abnormal value; judging the type of the evaluation index; and carrying out non-dimensionalization processing on the evaluation data according to different sections of the index types and the corresponding evaluation data sequences. The index evaluation data after dimensionless processing of the invention meets the condition that positive and negative indexes are unified into the same type and the minimum value is zero, and the change trend of the processed index data can be well matched with the original data; the method has the advantages that key information such as difference and variability between data before and after processing is better kept, the relative distance between samples is kept unchanged, and the balance of non-dimensionalized data distribution can be improved compared with the existing method.
Description
Technical Field
The invention relates to the field of comprehensive evaluation, in particular to an index dimensionless processing method for comprehensive evaluation.
Background
In order to improve the prediction and evaluation accuracy and facilitate the work of technicians, the relevant models need to be comprehensively evaluated, that is, information of a plurality of evaluation indexes of the evaluated models is mapped to one comprehensive index. And selecting an effective model from the multiple models so as to obtain a more accurate prediction or evaluation result.
In comprehensive evaluation, the unit of some indexes is length, some indexes are price, some indexes are area, some indexes are volume and the like, which are unit differences; some data are measured in thousands, and some data are measured in thousands, which is the difference of magnitude; some indexes are better if the data are larger, and are called as positive indexes, and some indexes are better if the data are smaller, and are called as reverse indexes; some indicators are best in a certain interval or a certain fixed value, called moderate indicators, which is a trend difference. The difference of different index data is mainly reflected in three aspects of unit, order and trend. In the data preprocessing of the comprehensive evaluation, it is generally necessary to perform non-dimensionalization processing (also referred to as normalization, or standardization of data) on the data in order to eliminate the differences in the three aspects and to achieve homounitization, homonymization, and homotrenization.
At present, common non-dimensionalization processing methods include sum-normalized processing, standard deviation processing, extreme value method, efficacy coefficient method, and the like. The sum standardization processing method and the standard deviation processing method cannot unify the indexes into positive indexes or negative indexes of the same type; the indexes subjected to non-dimensionalization by the extreme method can be unified into positive and negative same types, but are not applicable to the model evaluation method to be evaluated based on Dempster rule in particular because the minimum value is zero; the power factor method can avoid the situation that the minimum value is zero by adjusting the coefficient value, but cannot unify the positive and negative indexes.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides an index dimensionless processing method for comprehensive evaluation, which meets the requirements that positive and negative indexes are unified into the same type and the condition that the minimum value is zero is avoided, the change trend of evaluation data after dimensionless processing can be well matched with original evaluation data, key information such as difference, variability and the like between data before and after processing is kept as far as possible, and the relative distance between samples is kept unchanged.
To achieve the above object, the present application provides an index dimensionless processing method for comprehensive evaluation, comprising:
an index dimensionless processing method for comprehensive evaluation comprises the following steps:
s10, constructing original evaluation data sequences of different models to be evaluated according to evaluation indexes;
s20, segmenting the original evaluation data sequence corresponding to each evaluation index according to the abnormal value and the non-abnormal value;
optionally, the abnormal value is an extreme value deviating from most of the original evaluation data;
sequencing the original evaluation data of the n models to be evaluated by the evaluation index j from small to large, and recording as: { y1j,y2j,…,ynj};
Taking the median of the sorted evaluation data as a reference point, and identifying abnormal values;
the sorted evaluation data { y1j,y2j,…,ynjThe dividing structure is divided into three sections,
section one: for t identified between the minimum and median1An abnormal value of { y'1j,y'2j,…,y't1j},
A second stage: is t between the median and the maximum2An abnormal value, denoted as { y1j,y″2j,…,y″t2j},
Said t is1+t2+t3=n;
Setting the value range of the non-dimensionalized result of the evaluation data of different segments, wherein the value range of the non-dimensionalized result of the evaluation data of the segment I is set asThe value range of the non-dimensionalized result of the evaluation data of the second segment is set as The value range of the evaluation data dimensionless result of the third section is set as
S30, judging the type of the evaluation index,
optionally, the types of the evaluation indexes are divided into a positive index and a negative index;
s40, carrying out non-dimensionalization processing on the evaluation data according to the index type and different sections of the evaluation data sequence corresponding to the index type;
optionally, when the evaluation index j is the positive index, the piecewise non-dimensionalization processing expression is:
optionally, when the evaluation index j is the inverse index, the piecewise non-dimensionalization processing expression is:
wherein alpha is a segment weight coefficient, and the value range of alpha is 0< alpha < 1.
The application provides an index dimensionless processing method for comprehensive evaluation, which comprises the steps of firstly constructing an original evaluation data sequence of evaluation indexes on different models to be evaluated; segmenting the original evaluation data sequence corresponding to each evaluation index according to the abnormal value and the non-abnormal value; judging the type of the evaluation index; and carrying out non-dimensionalization processing on the evaluation data according to different sections of the index types and the corresponding evaluation data sequences. The index data after dimensionless processing of the invention meets the condition that positive and negative indexes are unified into the same type and the minimum value is zero, and the change trend of the processed index data can be well matched with the original data; the method has the advantages that key information such as difference and variability between data before and after processing is better reserved, the relative distance between samples is kept unchanged, and compared with the existing method, the method can improve the distribution equilibrium of non-dimensionalized data.
Drawings
In order to more clearly explain the technical solution of the present application, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious to those skilled in the art that other drawings can be obtained according to the drawings without any creative effort.
FIG. 1 is a flowchart illustrating a non-dimensionalization processing method for comprehensive evaluation according to the present invention.
Detailed Description
Referring to fig. 1, it is a working flow chart of an index dimensionless processing method for comprehensive evaluation according to the present invention;
as can be seen from fig. 1, a method for dimensionless index processing for comprehensive evaluation includes the following steps:
s10, constructing original evaluation data sequences of different models to be evaluated according to evaluation indexes;
s20, segmenting the original evaluation data sequence corresponding to each evaluation index according to the abnormal value and the non-abnormal value;
optionally, the abnormal value is an extreme value deviating from most of the original evaluation data;
sequencing the original evaluation data of the n models to be evaluated by the evaluation index j from small to large, and recording as: { y1j,y2j,…,ynj};
Taking the median of the sorted evaluation data as a reference point, and identifying abnormal values;
further, with the median of the sorted evaluation data as a reference point, the identifying an abnormal value includes:
the median is noted asSequentially calculating the distance d of the data on both sides of the median from the median, taking the maximum endpoint value and the minimum endpoint value as an example, and obtaining the following formula:
respectively calculate d→And d←If d is→>1, eliminating the end point value in the direction of the maximum value; if d is←>1, eliminating the end point value in the direction of the minimum value; repeating the step until d→And d←Are not greater than 1;
the sorted evaluation data { y1j,y2j,…,ynjThe dividing structure is divided into three sections,
section one: for t identified between the minimum and median1An abnormal value of { y'1j,y'2j,…,y't1j};
A second stage: is t between the median and the maximum2An abnormal value, denoted as { y1j,y″2j,…,y″t2j};
Said t is1+t2+t3=n;
Setting the value range of the non-dimensionalized result of the evaluation data of different segments, wherein the value range of the non-dimensionalized result of the evaluation data of the segment I is set asThe value range of the non-dimensionalized result of the evaluation data of the second segment is set as The value range of the evaluation data dimensionless result of the third section is set as
S30, judging the type of the evaluation index,
optionally, the types of the evaluation indexes are divided into a positive index and a negative index;
s40, carrying out non-dimensionalization processing on the evaluation data according to the index type and different sections of the evaluation data sequence corresponding to the index type;
optionally, when the evaluation index j is the positive index, the piecewise non-dimensionalization processing expression is:
optionally, when the evaluation index j is the inverse index, the piecewise non-dimensionalization processing expression is:
wherein alpha is a segment weight coefficient, and the value range of alpha is 0< alpha < 1.
While there have been shown and described what are at present considered the fundamental principles and essential features of the application, and advantages thereof, it will be apparent to those skilled in the art that the application is not limited to the details of the foregoing exemplary embodiments, but is capable of other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
Claims (5)
1. A method for dimensionless processing of an index for comprehensive evaluation, the method comprising:
constructing a data sequence of the original evaluation of different models to be evaluated by the evaluation index;
segmenting the original evaluation data sequence corresponding to the evaluation index according to the abnormal value and the non-abnormal value;
judging the type of the evaluation index;
and respectively carrying out non-dimensionalization processing on the evaluation data according to different sections of the index types and the evaluation data sequences corresponding to the index types.
2. The index dimensionless processing method for comprehensive evaluation according to claim 1, wherein segmenting the original evaluation data sequence corresponding to each evaluation index by an abnormal value and a non-abnormal value includes:
the abnormal value is an extreme value deviating from most of original evaluation data;
sequencing the original evaluation data of the n models to be evaluated by the evaluation index j from small to large, and recording as: { y1j,y2j,...,ynj};
Taking the median of the sorted evaluation data as a reference point, and identifying abnormal values;
the sorted evaluation data { y1j,y2j,...,ynjThe dividing structure is divided into three sections,
Said t is1+t2+t3=n;
Setting the value range of the non-dimensionalized result of the evaluation data of different segments, which is characterized in that the value range of the non-dimensionalized result of the evaluation data of the segment I is set asThe value range of the non-dimensionalized result of the evaluation data of the second segment is set asThe value range of the evaluation data dimensionless result of the third section is set as
3. The index dimensionless processing method for comprehensive evaluation according to claim 1, wherein the judging the type of the evaluation index includes:
positive and negative indicators.
4. The method according to claim 1, wherein the non-dimensionalizing the evaluation data according to different segments of the evaluation data sequence corresponding to the index type includes:
when the evaluation index j is the positive index, the piecewise dimensionless processing expression is as follows:
wherein alpha is a segment weight coefficient, and the value range of alpha is more than 0 and less than 1.
5. The method according to claim 1, wherein the non-dimensionalizing the evaluation data according to different segments of the evaluation data sequence corresponding to the index type includes:
when the evaluation index j is the inverse index, the piecewise dimensionless processing expression is as follows:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110845817.9A CN113506021A (en) | 2021-07-26 | 2021-07-26 | Index dimensionless processing method for comprehensive evaluation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110845817.9A CN113506021A (en) | 2021-07-26 | 2021-07-26 | Index dimensionless processing method for comprehensive evaluation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113506021A true CN113506021A (en) | 2021-10-15 |
Family
ID=78014602
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110845817.9A Pending CN113506021A (en) | 2021-07-26 | 2021-07-26 | Index dimensionless processing method for comprehensive evaluation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113506021A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105678453A (en) * | 2016-01-05 | 2016-06-15 | 国网山东省电力公司泰安供电公司 | Multi-dimension evaluation method on the basis of power grid index evaluation system |
CN108830506A (en) * | 2018-06-28 | 2018-11-16 | 国网山东省电力公司泰安供电公司 | Appraisal procedure, device and the realization device of city net power supply capacity |
CN112700027A (en) * | 2019-10-22 | 2021-04-23 | 中国科学院广州能源研究所 | Multi-precision index comprehensive evaluation method and system for electric power prediction model |
-
2021
- 2021-07-26 CN CN202110845817.9A patent/CN113506021A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105678453A (en) * | 2016-01-05 | 2016-06-15 | 国网山东省电力公司泰安供电公司 | Multi-dimension evaluation method on the basis of power grid index evaluation system |
CN108830506A (en) * | 2018-06-28 | 2018-11-16 | 国网山东省电力公司泰安供电公司 | Appraisal procedure, device and the realization device of city net power supply capacity |
CN112700027A (en) * | 2019-10-22 | 2021-04-23 | 中国科学院广州能源研究所 | Multi-precision index comprehensive evaluation method and system for electric power prediction model |
Non-Patent Citations (1)
Title |
---|
李伟伟 等: "综合评价中异常值的识别及无量纲化处理方法", 《运筹与管理》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110796173B (en) | Load curve morphology clustering algorithm based on improved kmeans | |
US20110238606A1 (en) | Kernel regression system, method, and program | |
CN110610431A (en) | Intelligent claim settlement method and intelligent claim settlement system based on big data | |
CN111079941B (en) | Credit information processing method, credit information processing system, terminal and storage medium | |
US10162867B2 (en) | Low memory sampling-based estimation of distinct elements and deduplication | |
CN114912720A (en) | Memory network-based power load prediction method, device, terminal and storage medium | |
WO2017073446A1 (en) | Device, method, and program for calculating explanatory variable value | |
CN110955862B (en) | Evaluation method and device for equipment model trend similarity | |
CN113506021A (en) | Index dimensionless processing method for comprehensive evaluation | |
CN112257958A (en) | Power saturation load prediction method and device | |
CN110751400A (en) | Risk assessment method and device | |
CN115081515A (en) | Energy efficiency evaluation model construction method and device, terminal and storage medium | |
CN114048776A (en) | Bridge damage identification method based on damage diagnosis index nonlinear narrow-area features | |
CN112884301A (en) | Method, equipment and computer storage medium for enterprise risk analysis | |
CN112348584A (en) | Vehicle estimation method, device and equipment | |
CN116843368B (en) | Marketing data processing method based on ARMA model | |
CN111753723A (en) | Fingerprint identification method and device based on density calibration | |
CN114386196B (en) | Method for evaluating mechanical property prediction accuracy of plate strip | |
CN113345538B (en) | Material performance prediction method based on dynamic selection training set | |
JP2018151913A (en) | Information processing system, information processing method, and program | |
WO2022156743A1 (en) | Feature construction method and apparatus, model training method and apparatus, and device and medium | |
WO2018009202A1 (en) | Continuous tissue analysis scoring scheme based on cell classifications | |
JP2004227279A (en) | Failure cause diagnosing method using mahalanobis distance and program | |
CN113590925A (en) | User determination method, device, equipment and computer storage medium | |
CN115034529A (en) | Training method, device and equipment of state recognition model and readable storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211015 |
|
RJ01 | Rejection of invention patent application after publication |