| Abstract [eng] |
The variety of the artificial intelligence and machine learning methods are applied for data analysis in various areas, including the data-rich healthcare domain. However, aiming to improve health care efficiency and use the captured information to improve treatment methods is often hampered by poor quality of medical data collections, as high percent of health data are unstructured and preserved in different systems and formats. In addition, it is not always agreed which methods of artificial intelligence and machine learning perform better in different problem areas, and which computer tools could make their application more convenient and flexible. The chapter provides essential characteristics of methods, traditionally applied in statistics, such as regression analysis, as well as their advanced modifications of logit, probit models, K-means, and Neural networks. The performance of the methods, their analytical power and relevance to the healthcare application domain is illustrated by brief experimental computations for investigation of stroke patient database with the help of several readily available software tools, such as MS Excel, Statistica, Matlab, Google BigQuery ML. |
