Research On Competence Modeling Automatically In The Context Of Online Innovation Outsourcing

As early as the year 2003,Professor Henry Chesbrough from UC Berkley proposed the perception of open innovation.As networking technologies develop fast,network-oriented innovation outsourcing service gradually becomes the important usage of open innovation.However,there are bottlenecks of the development of this service model.How to match tasks and talents based on a sea of information in an accurate and fast manner is one of the most urgent problems.The problem of achieving an efficient match online comes down to analyzing the competence information of online talents and the requirement of tasks automatically.Our work starts from the challenges facing online innovation outsourcing service,and this research includes five aspects:(1)Major-Course Association Knowledge Base Construction MethodsFirstly we used online data collecting techniques and nonstructural big data processing methods to research on the nonstructural data of education programs of universities.We used text analysis methods to build multi-system,multi-layer and multi-structural major-course association knowledge base.This serves as the basic knowledge base for inferring and building online talent knowledge structure model.Furthermore,we carried out a series of experiments to prove the accuracy and generality of this major-course association knowledge base.(2)Methods to Build Online Talent Knowledge Structure Model AutomaticallyFirstly,based on major-course association knowledge base,we designed a proper storage model for talent knowledge structure information and built human-computer interaction mechanism based on backend knowledge base,and then built the process of knowledge structure efficient read and write process.In the data processing during human-computer interaction,we designed a proper machine learning mechanism to maintain and update backend database.Furthermore,based on interaction behavior and information in knowledge base,we designed a way to automatically build online talent knowledge structure model.By combining experiments and case studies,we proved the accuracy and completeness of human-computer-interaction-based talent knowledge structure model.(3)Major-Skill Association Knowledge Base Construction MethodsBy analyzing the education programs' objectives and schema of universities,we applied text clustering and statistic methods,summarized and unified the skill requirement of different majors and built major-skill association knowledge base.We quantified the importance of different skills in different areas based on our statistic results,which strongly supported building online talent skill model.Finally,we carried out experiments to prove that the major-skill association knowledge base which was built upon area specific characteristics.(4)Methods to Build Online Talent Skill Structure Model AutomaticallyFirstly,based on our study of major-skill association knowledge base,we applied theories of knowledge expression and to design a layered talent skill structure storage model.Secondly,we studied talent education and experience oriented model analysis methods.Education oriented analysis was mapped to the associative relationships in backend database,while experience oriented analysis classified areas based on text classification methods.Based on these results,we built a human-computer interaction mechanism dependent on backend database and achieved efficient input of innovation skill structure information.Meanwhile,we designed proper machine learning mechanism to maintain and update backend knowledge base.We then achieved to automatically build talent skill structure model.We carried out experiments and expert evaluationmethods to evaluate theaccuracy and effectiveness of the skill structure model we created in this chapter.(5)Methods to Build Online Innovation Task Competence Requirement Model AutomaticallyFirstly,based on the knowledge base built by our study,we built a descriptive model of innovation tasks and analyzed the tasks by applying inference principles and text analysis techniques.We quantified knowledge requirement of innovation tasks by creating a vector representing course requirement.We also used task-major similarity to analyze the knowledge background related to online innovation tasks and obtained corresponding skills and weights of innovation tasks from major-skill association knowledge base and then created skill requirement model of tasks.We used case studies and experiment evaluation to demonstrate the effectiveness and accuracy of this automatic model building method.By the five chapters above,this research conducted innovations as follows:(1)We designed analysis methods oriented to multi-system,multi-layer and multi-structural education programs of universities for the first time,and introduced the characteristics of specific areas based on text clustering.Then we established quantitative and integrated major-skill association knowledge base and major-skill association knowledge base.The existing research focused on specific major areas and qualitative analysis.This research innovated in method and application,which preceded traditional methods.(2)We proposed automatic analysis method oriented to talent education background for the first time,in order to build talent knowledge structure model by human-computer interaction automatically.The existing talent knowledge structure modeling methods were based on subjective collection and manual building.This research innovated in online automatic modeling methods based on human-computer interaction,which showed the effectiveness on building online talent knowledge structure model.(3)We proposed automatic analysis method oriented to talent education background and practice information for the first time,in order to build quantitative talent skill structure model automatically.The traditional methods to analyze talent skill were based on were based on qualitative analysis and manual building.This research innovated in quantitative analysis and online automatic modeling methods,which preceded traditional methods in accuracy and efficiency.What's more,this research considered integrated effect of talent education background and practice information.(4)We quantified and built online innovation task competence requirement model automatically.This research advanced from tranditional manual qualitative analysis to automated quantitative analysis,and laid a basis for task-talent matching in the context of online innovation outsourcing.Our study proposed methods to build online outsourcing talent competence model and task competence requirement model automatically,which services as a solid foundation for efficient match of online outsourcing tasks and talents.Our studies can help improve the healthy development of online outsourcing innovation which is an emerging service model.