The paper presents a data mining approach as a possible solution for a medical problem of detecting bacteria proliferation syndrome in the small intestine by analysing glucose test results and patient survey data about intestinal tract. The proposed approach initially determines groups of similar objects - clusters, by analysing short time series given by the glucose test results; and then merges it with patient survey data having only most informative attributes. The obtained data set was used to nd relationships between clustering results and the descriptive parameters using classi cation algorithms. The obtained classi ers and patient self-assessment data served as a basis to determine whether the patient has to be assigned an advanced bacteria proliferation syndrome test in the small intestine or not. The article presents analysis of the acquired classi er evaluation results, as well as provides practical recommendations for medical experts that would use the proposed approach to detect bacteria proliferation syndrome in the small intestin.