A Multi-Stage Curriculum Development Model to Address Knowledge Gaps Between Academia and Industry

This innovate practice research is a work in Progress. Curriculum for technology programs must remain current with the recent developments and present needs of industry to ensure their continued success. Identifying and incorporating the newer developments requires a knowledge of those things that are important to the commercial users of technology. We have proposed and are testing a methodology that can easily allow educational programs to remain current using natural language processing followed up with confirmatory surveys. Our research used employment postings and website descriptions of ongoing research/ product development from technology companies to identify current interests. This data provided a window into where companies have invested their resources for future growth, which we used to understand where their technology needs were. The findings will then be used to determine what changes a program should make in their curriculum to address these needs. The information from these different sources was collected from the websites of selected technology companies as well as from technology related job listings. We then programmatically determined the ten most frequent words found in each subsection of the job listings search parameters. A correlation analysis for each word was obtained to give a context to the words in the resulting corpora. Our next step in developing this process was to extract more focused topics in our data from the correlations using focus topics extracted from the information available in the technology companies' sites. The process used to focus the topics in our data was started by capturing research and services related to health informatics and bioscience informatics. These research and services were obtained from a sample of 34 local and 39 national companies. The University of Massachusetts' (Amherst) free and open source machine learning toolkit, MALLET, was used to discover the topics present in the datasets via Latent Dirichlet Allocation (LDA) and then compare and interpret each topic present in the job posting frequent word correlations. This allowed us to determine more specific informatic subjects of interest along with the importance of these subjects to industry.Immediately following the analysis, we will incorporate our results into a survey, which will be sent to the companies that we identified in our data search. This will be used to validate our findings as well as obtain more specific information on their needs related to the curriculum. This second component will allow us to obtain more actionable information for making and incorporating new information and materials into the curriculum.