In this paper we describe the process of discovering underlying knowledge in a set of isokinetic tests (continuous time series) using data mining techniques. The methods used are based on the discovery of sequential patterns in time series and the search for similarities and differences among exercises. They were applied to the processed information in order to characterise injuries and discover reference models specific to populations. The discovered knowledge was evaluated against the expertise of a physician specialised in isokinetic techniques and applied in the I4 project (Intelligent Interpretation of Isokinetic Information)¹.