Projection pursuit dynamic clustering decision-making method for research and development project

Abstract

The invention discloses a projection pursuit dynamic clustering decision-making method for a research and development project.The method comprises the following steps of: firstly, acquiring a normalized score of each evaluation index of research and development projects of a success class, a postpone class and a failure class which are implemented and completed by a manufacturing enterprise through an expert scoring method, and establishing a K-DCPP decision-making technology model according to the normalized scores. The expert subjective initiative and K-DCPP objectivity are effectively combined, and a more reasonable, reliable, effective and objective decision result can be obtained.

Description

technical field [0001] The invention relates to a dynamic clustering decision-making method for research and development project projection pursuit. Background technique [0002] In the R&D practice of innovative projects of enterprises and scientific research institutes, there are not only many excellent R&D projects with great potential, but also many R&D projects with unqualified technology, unclear application prospects, inconsistent with macro policies, and insufficient investment in R&D technology. However, if these projects continue to be implemented, it will cause greater waste of financial, material and human resources. [0003] Therefore, for government science and technology (innovation) management departments and R&D enterprises, how to reasonably, reliably and correctly evaluate and optimize ongoing R&D projects, so as to select excellent R&D projects, and continue to increase these excellent R&D projects Investment efforts should be made to complete the pilot ...

AI Technical Summary

Problems solved by technology

[0005] 1) The traditional technical method uses fewer "teacher value" samples (usually a dozen to dozens) for the suspension decision-making, and the evaluation index data of the samples do not obey the normal distribution. The accuracy, rationality, and Reliability is difficult to guarantee

[0006] 2) Traditional technical methods cannot determine the attributes of R&D projects, and "ambiguous" results are likely to appear, and the judgment results are inaccurate

[0007] 3) The traditional technical methods do not consider or obtain the different weights of each evaluation index, so that equal weights are used or the importance of the evaluation index cannot be identified and sorted, and the index evaluation is unscientific

Images