基于高维稀疏数据分析的制造型企业研发投入对企业绩效的影响分析

Impact of R&D Input of Manufacturing Enterprises on Enterprise Performance Based on High Dimensional Sparse Data Analysis

影响企业绩效和科研成果的变量较多,采用定量分析方法对企业研发投入与绩效关系分析的局限性较大。探索一种对大量因素同时进行定量分析的方法,基于沪深两市327家制造企业2011—2016年数据,尝试采用高维稀疏数据分析方法,以"研发投入增长率>0和营业收入增长率>0和人均营业收入增长率>0"为目标,对影响目标实现的10类74个变量进行定量分析,得到实现目标的定量优化规则:当期和上期单位研发人员研发投入、研发人员投入增长率、盈利能力、产品市场竞争度5个变量处于一定范围内时,目标达成概率有较大幅度提升。研究结果可以用于衡量企业研发水平是否达到规模边界,又可作为企业研发投入行为对企业绩效结果的预判。

There are many variables that affect enterprise performance and scientific research results, so the quantitative analysis method has great limitations on the relationship between R＆D investment and performance. This paper explores a method for simultaneous quantitative analysis of a large number of factors. Based on the 2011 to 2016 data of 327 manufacturing enterprises in Shanghai and Shenzhen stock markets, this paper attempts to use the method of high- dimensional sparse data analysis, aiming at ＂ R＆D investment growth rate 〉 0 ＆ operating income growth rate 〉 0 ＆ per capita operating income growth rate 〉 0 ＂is the goal. Quantitative analysis of 74 variables affecting the achievement of the goal is carried out, and the quantitative optimization rule for achieving the goal is obtained, that is, when the five variables of the current and the last unit R＆D personnel R＆D investment, R＆D personnel input growth rate, profitability, and product market competition in a certain range, the probability of achieving goals has a large increase. The research results can be used to measure whether the R＆D level of the enterprise reaches the scale boundary, and also can be used as the pre-judgment of the enterprise＇s R＆D investment behavior to the enterprise performance result.