Enhancing the pattern recognition capacity of machine learning techniques: The importance of feature positioning

We design several algorithms representing evaluation processes of different complexity, ranging from basic environments based on a predetermined number features to complex structures involving alternatives defined through decision trees whose nodes is determined by the cardinality respective power sets. The sequential structure these builds information retrieval behavior users in online search environments. generate two strings data, namely, numerical evaluations determining and subsequent choices made latter. way output obtained inputted within vectors summarizing complexity conditions capacity machine learning techniques categorize them correctly. main purpose research illustrate numerically results. First, correctly even if their characteristic are presented that prevents identification using standard statistical techniques. Second, accuracy categorization capacities can be substantially enhanced describing required implement analyses. perform battery tests demonstrate analyze Their applicability classification prediction problems medical environments, particularly those constrained quality data available, emphasized. • defining complexity. positioning process distorted difficult identification. Categorization outperforms A technique performed validate emphasize problems.