The top four genes with a high absolute value of coefficient factor, namely, 2720 (GLB1), 3074 (HEXB), 6677 (SPAM1
), and 2799 (GNS), are selected to represent the genes for this canonical variable (Figure 4(b), Table 1).
This stream (Spam1 for short) is generated from the Spam Assassin stream by sampling a certain number ofinstances in the Spam class.
The Spam Assassin stream, Spam1 stream, and Spam2 stream use the interleaved chunk procedure by collecting 300 instances as a textual chunk for one time stamp in a proper sequence.
The average accuracy in the Spam1 stream is shown in Table 2.
Meanwhile, in the Spam1 stream, CFIM, AUE-RF, and LB-RF achieve the highest level of kappa statistic.