Research And Application Of The Talent Market Demand Forecasting Model

Talent resource is the first resource. Talent, which has fundamental, decisive, andstrategic effect on social and economic development, is the most important and active factor inproductive forces. With the development of economic globalization in recent years, globalindustries are accelerating transformation and optimization. At the same time, the socialeconomy also put forward more and more requirements for talents to adapt to the change anddevelopment of economic. Therefore, it has been the most urgent need for job seekers,enterprise, regional and even global organizations to forecast the requirements of talent marketreasonably and effectively.Online recruitment, which has the features of plenty of talent information, wide coverage,less investment and return for more, good recruitment performance, automation for talentmatching, and so on, is becoming the preferred approach of job hunting and recruitment formore and more job seekers and recruiters.This paper conducts a preliminary forecasting for the talent market requirements of Hprovince by collecting, analyzing, statistical computing, and data mining the database from thean online recruitment website. The detailed research is as follows:1.It analyses the requirement from the online recruitment of H province, clarifies theresearch motivation. It studies the multidimensional analysis technology for data description,data prediction based on probability and statistics, and prediction based on data mining.2. It builds a multidimensional analysis model to find the basic features of the onlinerecruitment web market. The analysis for the basic situation of provincial talent market duringthe last five years from the aspects of job seekers’ sex structure, age structure, educationbackground (degree) structure, work experience structure, major structure, and positionstructure finds some useful results. For example, the recruitment rate that employers do notrestrict the talents’ gender accounts for about74%and the recruitment percentage thatemployers do not restrict the talents’ age accounts for about43%. The recruitment withrequirement that job seekers’ age is between eighteen and thirty accounts for about22%. Theeducational requirement of provincial talents relatively focuses on undergraduate and collegedegree. About67%employers recruit without the requirement for work experience. The job seekers’ major distribution is concentrated on electronic information, management,engineering, economy, enginery, civil, etc. The requirement for sales personnel have beenoccupied the high position about29%and next is the requirement for enginery and civiltalents in the distribution of position requirement.3. It chooses grey prediction model, which needs less historical data with high precision,to make a short-term prediction for talent quantity demand. The total demand of talents isabout168600people in2013.4. It combines decision tree algorithm with time series model to forecast the talentstructure requirement. Firstly, in order to make a forecast for talent structure requirement, itdivides the talent structure into natural structure, social structure and economic structure.Secondly, it employs Microsoft decision trees algorithm and moving average algorithm tomake a short-term prediction for talent requirement structure of provincial talent market. In thetalent natural structure requirement, it finds that what have an intense influence on talentapplication is gender structure, weight structure (shape demands), age structure andemployment status structure. In the talent social structure requirement, what have animportance influence on talent application are work experience structure, educationbackground structure, and mandarin structure. In the talent economic structure requirement,position categories and major categories have a significant influence on talent application.5. It proposals a phasic conditional probability model based on the idea ofstructuralization. The recruitment market presents phasic pattern from the overall perspective.By the knowledge so far, it has not find the current statistic method and data mining methodconsider the pattern. It arguments the stable periodic and non stable periodic time series modelrespectively based on the phases at the end of the part of the total cumulative amount ofconditional probability distribution model. It estimates the phase parameters of the forecastphase by the maximum likelihood parameter estimation method, and makes a predictionfinally.In summary, by this research of talent market requirement, this paper produces thefeedback of talent requirements information to job seekers. It helps job seekers get theaccurate employment demand forecasting information so that they can have an insight into the future employment situation more clearly, they can find their ideal jobs quickly, they canchoose a proper major, therefore, they can better adapt to the future employment status.