3D star coordinate projection

Kandogan introduced the concept of star coordinate projection in 2000 [1]. To illustrate the idea behind 2DSCP, consider an example data point of dimension 8 given by 0.98]. 0.95 0.06 0.869 0.01 0.43 37 . 0 [ = A The dimensional anchors for 2DSCP are shown in Fig. 1. The dimensional anchors multiplied by the dimension values for data point ‘A’ are shown in Fig. 2. As shown in Fig. 2, each of the dimensional anchors is scaled to the dimensional values of data point ‘A’. The scaled dimensional anchors for data point ‘A’ are connected in order as shown in Fig. 3. The point of destination after covering all the dimensional anchors is the point of representation for the data point ‘A’. As a result, the 8 dimensional data point is now represented in two dimensions. The same process is repeated for all the data points. The dimensional anchors as shown in Fig. 1 may be arranged in multiple ways there by offering multiple ways to project the data. Also, in the above example, it is assumed that all the dimensional anchors are of equal length as shown in Fig. 1. Kandogan proposes use of random lengths of dimensional anchors to assign importance [1]. The different arrangements coupled with different length assignments of dimensional anchors are explored until a pattern is observed in the data. The number of arrangements that need to be explored is not known and hence, as will be shown later, the automated method that provides possible best projections among the possible projections is extremely important for SCP based projections. Fig. 1. 2D Star Coordinate