L. Pompilio, “Analysis of 10 CFR Part 810 General Authorization Data on Assistance to Foreign Atomic Energy Activities”, M.S. Thesis, Nuclear Engineering, Texas A&M University, College Station, TX (2017).
Under contractual agreement, Texas A&M has received and converted the Part 810 General Authorization data to a searchable format using Optical Character Recognition (OCR). In conjunction with Argonne National Laboratory (ANL), applicable categories were chosen to group various report elements together in an Excel Spreadsheet. Example categories include, country of recipient, transfer dates, and 10 CFR 810.2. To expand, 10 CFR 810.2 is how the given authorization is tagged and includes Uranium conversion.
Texas A&M and ANL had the original goal of modeling the data, however, upon completion of the Excel spreadsheet it was determined that modeling options were not applicable, as the data is qualitative over quantitative. Other exploratory goals were determined and are as follows: Searching, Reporting, Analyzing, and Predicting. The Excel spreadsheet, as mentioned above, will implement searching; acting as a central repository to facilitate ad hoc searching. Reporting requirements have been established using the visual analytics software, Tableau, wherein general authorization trends were determined and can be subsequently monitored. The machine based learning algorithm, Apriori, has been used to determine data mining rules that can be implemented and modified to predict the occurrence of an item based on the occurrences of other items in the supplied item set.
The item set for the purposes of this paper are the categories in which the reports were tagged. Prediction has been carried out using Tableau’s forecast option, that will anticipate the number of general authorizations to be received by a given country based on prior requests. Excel modeling has also been utilized for this purpose. Paterva’s Maltego software has been utilized to search the internet and determine when a general authorization report is not received based upon news reports. The specific transform to be used is the phrase transform, wherein a key phrase or part thereof is entered and searched for on various websites. It is up to the discretion of the user to prioritize the given sources. Supplemental to the exploratory goals, four questions were posed for presentation to the Department of Energy (DOE). Using Tableau’s analytics software, it has been determined that there are no significant trends for non-specifically authorized destinations as pertaining to specific transfers. Additionally, enrichment has been found to be the most dominant sensitive nuclear technology transfer, while no alarming technology transfers were identified as changing over time to any given country. Certain U.S. companies, however, have been identified as entities that have only transferred to certain companies. The above provides a system of checks and balances for the Part 810 General Authorization Process.