ABSTRACT
A database on cyclic shear tests of unreinforced masonry walls with 288 entries has been assembled. The database contains data on stone, solid and hollow brick masonry and on masonry strengthened by different methods. The database is analysed using a data mining and machine learning software to find correlations using regression type of methods and construct decision trees. The most common experiment to observe behaviour of masonry under seismic loads is the cyclic shear test. In addition to providing insight into the behaviour of the wall in seismic conditions, the test can be used for assessing the tensile strength of masonry. The chapter presents the use of two types of algorithms: the first group are algorithms which rank the parameters according to their relevance, and in the second group there are algorithms for building actual models.
