TPAC - Technology Policy and Assessment Center at the Georgia Institute of Technology

Technology Opportunities Analysis (TOA) exploits electronic information resources to provide technology foresight. The TOA approach proceeds as follows:

TOA exploits electronic information resources for technology analyses. TOA helps monitor activities, forecast developments, and assess the potential implications of emerging technologies. The target of opportunity is the wealth of information compiled in large bibliographic databases. For instance, Medline has abstracted well over 10 million health-related journal articles and conference presentations. TOA adds value by processing the text and counting activity reflected in 100’s to 10,000’s of abstracts gathered on particular topics. Such analyses are called "bibliometrics," "knowledge discovery in databases," or "data mining." To facilitate these analyses we began developing TOA software at Georgia Tech’s Technology Policy and Assessment Center (TPAC) in 1993. The work has advanced in partnership with:

A number of sources have collectively provided over $2 million to develop TOA:

TOA software uses several approaches to ascertain patterns in a body of text records. It applies natural language processing to parse text into phrases and parts of speech for further analyses. It combines rules and elemental fuzzy logic to remove duplicate records, consolidate like terms, and so forth. Thesaurus and grouping capabilities allow examination of subsets of the set of records. Statistical analyses, particularly Principal Components Analysis (PCA), help find relationships among topics and performers. These mainly build on co-occurrence that is, certain terms tending to appear together in records is taken as evidence of possible relationship. Representation of the analytical findings is a critical element in TOA. The software enables key tabulations, trend plots, and relationship maps. Proprietary algorithms include multidimensional scaling and path erasing to generate sets of maps and selected innovation indicators. We continue to advance TOA approaches and applications. Interests include generation of particular technology development indicators, processing of foreign text, and automated abstracting.

TOA begins by tallying lists and counts for various fields over the record set. We often generate "Top 10" lists of research organizations, topics, and so on. Lists can then be compared to ascertain common and unique features. Figure 1 combines histograms on two organizations’ patent activity to show how their interests differ.

Co-occurrence matrices cross two lists (e.g., keywords by keywords, countries by keywords) to show concentrations of activity. Figure 2 illustrates a truncated cross-tabulation of Japanese companies by technology.

We use a proprietary PCA-based approach called "Principal Components Decomposition" to "bucket" subsets of the records in smaller, more accessible groups. Another variant of this can help identify frontier interests on "the fringe" of the main topics. Figure 3 clusters a search set of over 3,000 abstracts to facilitate cross-disciplinary exploration of related interests within the general domain of diesel engine R&D.

Using general co-occurrence information and another PCA tool, we can generate a family of maps relatively automatically. These include topic cluster maps, country/topic maps, and so forth. Figure 4 shows a map of leading "internet" topics (from a search in the INSPEC database)that also notes two research organizations publishing the most on each topic. This map facilitates understanding the scope of "internet" R&D. It can guide detailed probing by topic or organization.

Trends in activity are often of special interest. These typically begin with overall activity (e.g., patenting rate over time). Further trends may spotlight "hot" subareas. Figure 5 is somewhat more elaborate. It shows time slices for three variables (from back to front: number of publications, number of organizations publishing, and number of keywords found in those publications). An intriguing phenomenon is represented here. Publication rate plummeted in the 1990’s on ceramics for engines, rebounding slightly in 1993-95; however, the number of keywords skyrocketed in 1993-95. Content analyses then showed that the very nature of those keywords changed with tremendous expansion of research attention to production and materials issues. These signaled significant maturation of the underlying technology, warranting initiation of major new projects by TACOM.

TPAC is working to generate innovation indicators empirical measures that relate to key technology life cycle, contextual influence, and market potential concepts. Figure 6 shows one such indicator. It shows??

TOA adances as new applications suggest desirable software advances. Applications range widely:

Different users find value in different applications, for instance:

In sum, TOA supports a range of technology (and other topical) analyses that summarize vast amounts of information. Typically, we search on specific topics to compile electronic records reflecting publication, patent, project, and/or citation activity. We analyze these records and represent what we find via text, tables, and graphs. Those can serve such uses as:

Fuel cells
Knowledge Discovery in Databases and Data Mining
Internet Mapping Example
Publications and Projects

How can you explore TOA or TPAC interests further?

See below for example TOA analyses and further information. To explore having a TOA done for you, contact IISC (Intelligent Information Services Corporation). To explore obtaining TOAS software, contact Search Technology.

TPAC (Technology Policy and Assessment Center) is a Georgia Tech interdisciplinary research center. Founded in 1981, it reports to the School of Public Policy and is affiliated with the School of Industrial and Systems Engineering. TPAC faculty and students pursue various technology analyses. Particular strengths include R&D evaluation, technology foresight, and technology-based economic development analyses.

You are welcome to contact Alan Porter, Director of TPAC:
e-mail: alan.porter@isye.gatech.edu
phone: 404-894-2330
fax: 404-894-2301
mail: ISyE, Georgia Tech, Atlanta GA 30332-0205