Quantitative Analysis Of Redundant Labor Force Migration Based On Artificial Neural Network

The migration of rural redundant labor force to cities is the inevitable trend of industrialization and urbanization, an indispensable stage during the economy development, and the necessity of human society process.By the influence of such factors as socioeconomic patterns of different areas, forms of local organizations and population distribution, regional disequilibrium of rural labor force migration has become more and more serious. It will make great sense for finding out migration rules, making migration policies adapting to local circumstances, speeding up the migration of rural redundant labor force and taking advantage of human resource, in order to analyze the regional differences of rural redundant labor force migration and predict the future trend.Some conventional methods are widely used in quantitative analysis of labor force migration, including time sequence model, regression analysis model and econometric model etc. However, as a multivariable nonlinear system and because of the complicated correlation among migration factors, labor force migration modeling using existing methods is not fruitful as expectation: they are limited on both the number of research objects and prediction accuracy. The key problem addressed in this research is how to investigate the process of labor force migration quantitatively with artificial neural network (ANN).This paper focuses on the pure labor services exports among provinces in mainland. Most of them are redundant rural labor force from different regions, which have come into large amount in our country during the past few decades. We exploit the potential of ANN to model the labor force migration system. This research has three main goals. First, analyze the properties of migration factors among inter-province labor migration quantitatively. This includes exploring the interactions between migration factors and decisions and predicting the floating amount and direction of labor force. Second, provide reasonable policy suggestions according to the quantitative results. Third, propose a general framework based on ANN to model nonlinear systems as labor force migration. This paper has four main contributions. (1) Propose a quantitative method based on ANN to study how migration factors influence the decisions through static and dynamic elasticities. (2) To the best of our knowledge, this research is the first one to propose a referenced formula of migration strength. Based on migration strength, regional characteristics during migrations are investigated quantitatively. (3) Presents how to predict the migration amount and probability using ANN, which leads to better results than conventional methods in some aspects. (4) Introduces a new approach on the quantification of non-quantitative variables, which makes up for the deficiency of current data collecting method.All conclusions and suggestions in this paper are derived from quantitative results, which is the most important feature distinguishing this research from others. The major viewpoints .of elasticity analysis can be classified into the following four categories. (1) Conclusions same as the current ones. For example the decisive influence of wage level in ingoing place in contrast to the minor influence of the wage level in outgoing place, the restraining effect of migration distance, the influence of the average level of education in outgoing place and so on. (2) Conclusions different from the current ones. For example the negligible influence of the average agriculture acreage in outgoing place to the migration decisions. (3) New general conclusions. For example the minor influence of wage level in ingoing place, and the great promotion of nonagricultural population rate. (4) New special conclusions. For example, being affected by the current situation in our country, the quantity of employment in secondary industry prohibits migration noticeably, while the quantity of employment in tertiary industry boosts migration. Analysis of migration strength discovers notable regional effect in labor force floating.The quantitative research framework concluded from this article has good adaptability. Artificial neural network can be applied in different research topics of labor force migration, and some valuable results are expected.