Data Mining in the Analysis of Tree Harvester Performance Based on Automatically Collected Data

Data recorded automatically by harvesters are a promising and potentially very useful source of information for scientific analyses. Most researchers have used StanForD files this purpose, but these troublesome to obtain require some pre-processing. This study utilized new similar data: JDLink, cloud-based service, run the machine manufacturer, that stores data from sensors in real time. The vast amount such makes it hard comprehend handle efficiently. mining techniques assist finding trends patterns databases. Records two mid-sized working north-eastern Poland were analyzed using classical regression (linear logarithmic), cluster analysis (dendrograms k-means) Principal Component Analysis (PCA). Linear showed average tree size was variable having greatest effect on fuel consumption per cubic meter productivity, whereas hour also dependent, e.g., distance driven low gear or share time with high engine load. Results clustering PCA harder interpret. Dendrograms most dissimilar variables: total volume harvested day, day work revolutions minute (RPMs). K-means allowed us identify periods when specific clusters variables more prominent. results, despite explaining almost 90% variance, inconclusive between machines, and, therefore, need be scrutinized follow-up studies. Productivity values (avg. around 10 m3/h) rates (13.21 L/h, 1.335 L/m3 average) results reported other authors under comparable conditions. Some measures obtained include, (around 7 km day) proportion running low, medium load (34%, 39% 7%, respectively). assumption use without supplementing external sources, as little processing possible, which limited analytic methods unsupervised learning. Extending database studies will facilitate application supervised learning modeling prediction.