Variant of Data Particle Geometrical Divide for Imbalanced Data Sets Classification by the Example of Occupancy Detection

The history of gravitational classification started in 1977. Over the years, approaches have reached many extensions, which were adapted into different problems. This article is next stage research concerning algorithms creating data particles by their geometrical divide. In previous analyses it was established that Geometrical Divide (GD) method outperforms algorithm based on classes a compound 1 ÷ cardinality. occurs process balanced sets classification, class centroids are close to each other and groups objects, described labels, overlap. purpose examine efficiency unbalanced example real case-occupancy detecting. addition, paper, concept Unequal (UGD) developed. evaluation conducted 26 sets-16 with features Moons Circles 10 created occupancy set. experiment, GD its variant as well 1CT1P approach, compared. Each combined three particle mass determination algorithms-n-Mass Model (n-MM), Stochastic Learning Algorithm (SLA) Bath-update (BLA). k-fold cross validation method, precision, recall, F-measure, number used applied process. Obtained results showed methods divide outperform approach imbalanced classification. article’s conclusion describes observations indicates potential directions further development methods, concern through