Application of Data-mining Methods in the Assessment of Excessive Vibrations: Decision Trees, Basket-market Analysis

The article examines the issue of assessing the impact of above-limit vibrations on the health of employees in handling and dispatch warehouses (timber yards). The results are based on the arithmetic mean of multiple measurements of both hands of the chainsaw operator or the hands/whole body of the machinist, while taking care to maintain consistency. The measurement locations, measurement duration and selection of samples were chosen so that the result represented and described the employee's personal exposure. The main objective was to identify the relationship between the working activity of the studied forest administration employees, exposure to vibrations, and the actual accurately identified harm to the health status of the assessed persons. Basic methods of statistics and probability theory and methods of in-depth analysis (decision tree, association rules) were used in the evaluation. From the results, significant exceeding of the action and limit values of exposure to vibrations transmitted to the hands at the work position (chainsaw operator) was found. The decision tree showed that the age and experience in the employee's field have a major impact on the development of traumatic vasoneurosis (part of vibration disease occurring because of long-term local exposure to excessive vibrations). Through the analysis of association rules, dependencies were found between selected attributes that described each employee (age, experience, exposure, vibration acceleration, medical history). From the association rules, we obtained significant associations, for example, between a high value of vibration acceleration and numbness of the fingers, or back pain, tinnitus, etc. Also based on the analysis of the association rules, it was confirmed that old age or long experience in the field and exposure to vibrations over 4 hours/day significantly affect the occurrence of vascular changes in the fingers of the studied subjects.