Font Size: a A A

Assessment Of Innovation And Index Prediction Of High-Tech Development Zones In Inner Mongolia By Machine Learning

Posted on:2014-10-22Degree:DoctorType:Dissertation
Country:ChinaCandidate:R ZhangFull Text:PDF
GTID:1269330422968197Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
High-tech development zones in Inner Mongolia played an important role for thewestern development drive. Innovation capacities evaluation of zones helped a lot todeveloping programme for administrator. The programme could promote the develop-ment of zones deeply.Firstly assessment indexes was established for regarding enterprizes as innovationsubjects. Nine representative High-tech zones were selected out. The data in2010wasstandardized with logarithm function. Correlation was detected by rank correlation co-efcients. The selected indexes satisfied independence principle. Principal componentanalysis and factor analysis were applied to obtain order of innovation for nine zones.The factors influenced innovation capacity with a network approach. The greatvirtue of machine learning assessment was its objectivity. but its sample was an obsta-cle. Here AHP was used to help design sample. The weights were regarded as normalrandom variables. The value of AHP could be thought as mean and consistency ratioof judge matrix as variance. Random weights were generated by Monte Carlo. Thussample came out. The assessment values and orders were acquired by Random Forest,Stochastic Gradient Boosting, Support Vector Machine and etc six methods. Averag-ing the six evaluation values was the final order with respect to combination predictiontheory.Industrialization of new technology need a cycle time. Predicted GDP was morebetter than presented value to evaluate scientifically. GDP time series of JinChuan de-velopment zone were detected to be independent but not identical distribution by inde-pendence test of BDS, Box-Pierce and Ljung-Box. No conditional heteroscedasticityexisted. Normal distribution was reasonable. So Moving Averaging model could beconsidered. Correlation and partial correlation functions, AIC and BIC information cri-teria, independence of residuals were three tools for order determination. MA(4) waspicked lastly. GDP of2014was predicted to assess innovation capacity more objec-tively and more reasonably.Finally, Fostering high-tech industrial clusters, enhancing the capability of inde-pendent innovation and improving the scientific and technological innovation servicesystem were viewed as three development goals. Seven Strategic measures were pro- posed. Thus it would do a new leap in high-tech industry of Inner Mongolia Au-tonomous Region.
Keywords/Search Tags:High-tech developments zone, Innovation capacity, Comprehensive assess-ment, Random forest, Stochastic gradient boosting, Nonlinearity test
PDF Full Text Request
Related items