Mechanizing Hypothesis Formation: Principles and Case Studies

Mechanizing hypothesis formation is an approach to exploratory data analysis that uses a one-to-one relation between theoretical and observational statements to generate all interesting theoretical statements. The LISp-Miner system developed at the Prague University of Economics and Business includes several GUHA procedures for association rules,couples of association rules,action rules,histograms,couples of histograms,and patterns based on general contingency tables. Recent research results relevant to dealing with domain knowledge in data mining and its automation are also provided.

Format: Hardback
Length: 346 pages
Publication date: 20 October 2022
Publisher: Taylor & Francis Ltd

Mechanizing hypothesis formation is a crucial approach to exploratory data analysis, with its origins dating back to the 1960s. This innovative approach was sparked by the intriguing question, "Can computers formulate and verify scientific hypotheses?" The result of this development was a comprehensive theory of logic of discovery, encompassing both theoretical and observational calculi. These calculi are interconnected through statistical hypotheses tests, forming the foundation of this powerful methodology.

One notable tool within the logic of discovery is the GUHA method, which employs a one-to-one relationship between theoretical and observational statements to extract all interesting theoretical statements. This method then generates all relevant observational statements and verifies them using a given observational data set. The output of the procedure consists of all observational statements that are true within the provided data.

The LISp-Miner system, developed at the Prague University of Economics and Business, is a comprehensive platform that incorporates various GUHA procedures for dealing with association rules, couples of association rules, action rules, histograms, couples of histograms, and patterns based on general contingency tables. This system has been instrumental in achieving numerous results related to observational calculi and has been applied in real-world scenarios within data mining and business intelligence.

The book provides a concise yet comprehensive overview of the logic of discovery, offering valuable insights into this field. Numerous examples of applications of the GUHA procedures to solve practical problems in data mining and business intelligence are presented, showcasing the practical utility of this methodology. Additionally, an overview of recent research results pertaining to dealing with domain knowledge in data mining and its automation is provided, highlighting the ongoing advancements in this domain.

Furthermore, the book offers firsthand experiences with the implementation of the GUHA method in the Python language, making it accessible to a wide range of practitioners. Whether you are a data scientist, business analyst, or researcher, this book provides valuable guidance and practical techniques for automating hypothesis formation and extracting meaningful insights from data.

Weight: 830g
Dimension: 234 x 156 (mm)
ISBN-13: 9780367549800