Prediction of PWEB Unit and BRIAN Unit based on Cataloging Unit, Content, and Administrative Unit
Yanjuan Zhang, Jing Zhao, Panickos N. Palettas, Thomas J. Santner · 2001
and Subsite (96), the missing observations has the lowest counts (3). In BRIAN unit, Page again has the highest counts (175), followed by Subsite (127) and Site (90), the count of missing observations is 22, which is higher than that in PWEB unit. In Cataloging unit, Whole turns out to be the most coded values (246 counts), Analytic is in the second place (149), missing observations is the third (11) 2 and Defies classification has only 8 counts. In Content, the rank by frequency count is ordered as Bottom (186), Top (142), Middle (85), and missing observation (1). Finally, in Administrative unit, Top takes the largest proportion (206), followed by Bottom (107) and Middle (82), with the count of missing observations being 19. We noticed that the percentage of missing observations is so small, thus we decided to drop them all for further analysis. Clearly, for the remaining data, PWEB unit has three categories: Page, Site, and Subsite, which are the same for "BRIAN u