Automated Extraction and Classification of Slot Machine Requirements from Gaming Regulations

In well-regulated industries, important technical requirements can often be found in state and federal laws and regulations. This paper examines how natural language processing can be used during requirements analysis to analyze government regulations. The examples used in this paper are drawn from casino industry regulations for slot machine development, but are applicable to analyses of government regulations in other industries as well. More specifically, this paper analyzes South Dakota and Nevada regulations for slot machines and applies natural language processing to extract and analyze technical requirements derived from them using four techniques. First, key words and key phrases are drawn from the regulations using the Rapid Automatic Keyword Extraction algorithm so that they can be imported into a program glossary. Second, requirements are extracted from the regulations. Many of these requirements do not have the word “shall”, so a 12-rule transformation algorithm is used to convert the text into “shall” or “may” statements. Third, a Naive Bayes model is developed from the South Dakota regulations to predict which of the extracted Nevada requirements are functional, and which are not. Finally, a Dice similarity metric weighted with term frequency-inverse document frequency scores is used to identify related and equivalent requirements between the South Dakota and Nevada regulation sets.