基于机器学习的未来工业投资重点模型研究

Research on the Key Model of Future Industrial Investment Based on Machine Learning

摘要: 经济是国家的基础，经济安全是国家安全体系的重要组成部分，是国家安全的依托和基石。我国经济已由高速增长阶段转向高质量发展阶段，正处在转变发展方式、优化经济结构、转换增长动力的关键期。中国GDP由各行业GDP所组成，研究各行业相关性，分析中国未来在各行业的最优投资以及如何刺激就业已成为当前重要课题之一。根据历年来各行业投入、产出和就业基本情况等数据，并基于皮尔逊相关系数矩阵、线性回归分析、投入–产出模型等研究方法，对各行业的相关关系，投资与各行业GDP的相关关系进行了详细分析，建立基于历史数据的各行业GDP与刺激就业的模型，并使用拉格朗日乘数法对该模型进行了求解。首先搜集数据，对数据进行预处理，检查中国各行业历年来的GDP是否有缺失值、异常值等，并对数据分类汇总处理。其次利用可视化处理、皮尔逊相关系数矩阵和线性回归分析等方法，详细分析了各行业之间的相关关系以及各行业对中国经济的促进和制约。本文我们建立了投入–产出模型、CES生产函数、Solow增长模型和VAR模型，并运用中国2010年之前的数据作为训练集，预测了2011~2023年的投资和各行业GDP之间的关系。之后，根据预测结果，得出最优模型为投入–产出模型。最后，根据各行业的回归系数详细分析了投资与各行业GDP的相关关系。我们从一万亿投资资金出发，结合第二问的最优投资分配投入–产出模型，采用Cobb-Douglas生产函数，构建总GDP最大化的投资分配模型，运用拉格朗日乘数法对该模型进行求解，给出了在限制投资资金时的最优投资方案，并详细分析了限制三个行业时的最优投资方案。结合前面的分析和建模过程，提出了刺进就业和改善工作质量的就业最大化模型。并运用拉格朗日乘数法对该模型进行求解，得出提高就业率和工作质量的最优投资分配，并详细分析了限制三个行业时的最优投资方案。

Abstract: Economy is the foundation of a country, and economic security is an important component of the national security system, serving as the support and cornerstone of national security. China’s economy has shifted from a stage of high-speed growth to one of high-quality development. It is currently in a critical period of transforming the development mode, optimizing the economic structure and converting the growth drivers. China’s GDP is composed of the GDP of various industries. Studying the correlation among industries, analyzing the optimal investment in various industries in China in the future and how to stimulate employment has become one of the important current issues. Based on the data such as the basic situations of input, output and employment in various industries over the years, and based on research methods such as Pearson correlation coefficient matrix, linear regression analysis, and input-output model, a detailed analysis was conducted on the correlations of various industries and the correlations between investment and the GDP of various industries. A model of the GDP of various industries and employment stimulation based on historical data was established. The Lagrange multiplier method was used to solve the model. First, collect data, preprocess the data, check whether there are missing values, outliers, etc. in the GDP of various industries in China over the years, and classify, summarize and process the data. Secondly, by using methods such as visualization processing, Pearson correlation coefficient matrix and linear regression analysis, the correlations among various industries and the promotion and constraints of each industry on the Chinese economy were analyzed in detail. In this paper, we established the input-output model, the CES production function, the Solow growth model and the VAR model, and used the data of China before 2010 as the training set to predict the relationship between investment and the GDP of various industries from 2011 to 2023. Afterwards, based on the prediction results, the optimal model is derived as the input-output model. Finally, the correlation between investment and the GDP of each industry was analyzed in detail based on the regression coefficients of each industry. Starting from one trillion investment funds, combined with the input-output model of the optimal investment allocation in the second question, adopting the Cobb-Douglas production function, we constructed an investment allocation model that maximizes total GDP. We used the Lagrange multiplier method to solve this model and presented the optimal investment plan when limiting investment funds. The optimal investment plan when restricting three industries was analyzed in detail. Combined with the previous analysis and modeling process, an employment maximization model for penetrating employment and improving job quality was proposed. The Lagrange multiplier method was used to solve the model, obtaining the optimal investment allocation for improving the employment rate and job quality, and the optimal investment schemes when limiting three industries were analyzed in detail.