Then you'll need to used comparable corpora of English and Chinese. The Lancaster Corpus of Mandarin Chinese LCMC (which is freely available from the Oxford Text Archive) and the British English FLOB / the American English FROWN are ideal for your purpose.
"Raw frequency" in the LCMC query report is the same as "Frequency" in Babel, both referring to the actual number of occurrences in the corpus. "Frequency per 10K words" in the LCMC report refers to the "normalized frequency" in each 10,000 words so that different genres can be be compared with each (as different genres are of varying sizes).