Alternating Least Square Method for Decomposing Dance Golek Menak Tensor Data

Abstract

Golek Menak Dance rhythms a body structure that is motion and geometric patterns that can be presented in the form of a vector, matrix and tensor. In this study the data obtained from the catcher dance Kinect sensor in the form of Bio Vision Format Hierarchy (BVH), which is a geometric data consisting of coordinate geometry (x, y, z), time as well as reinforcing the position of the body (skeleton). The motion data is compiled into a data tensor matrix decomposition process is then performed for a feature that can be obtained will be processed in motion classification. Alternating Least Square (ALS) methods are used in the data decomposition tensor.