Co-occurrence data representing e.g. preferences and facts observed in a plurality of situations may be stored in a matrix as combinations of high-dimensional sparse vectors. The matrix may be called e.g. as an experience matrix. The data stored in the experience matrix may be subsequently utilized e.g. for predicting a preference of a user in a new situation. Co-occurrence data may be stored in the experience matrix may be updated by a method comprising determining a first word based on a state of a system and/or based on a physical quantity detected by a sensor, forming a first word group comprising the first word and a second word, associating the first word and the second word with a common sparse vector, associating the first word words with a first vector of a matrix, associating the second word with a second vector of the matrix, modifying the first vector of the matrix by adding contribution of the common sparse vector to the first vector of the matrix, and modifying the second vector of the matrix by adding contribution of the common sparse vector to the second vector of the matrix.