There was one that conceivably could have been a bad OCR attempt to read Chinese characters using a Latin alphabet which is weird, but some of them do also look like bad/non-standard Romanisation attempts (with an odd lack of spaces etc.). Odd because I'd think technology to accurately generate standard (pinyin) romanisation must be reasonably good by now.
The lack of spaces is somewhat normal if you assume that chinese doesn't use spaces when writing, like a name is just the glyphs combined together in one word, while in English normally there is a space between F+L.
The point of Romanisation is to make it possible to at least mentally turn it into "sounds" for those familiar with Latin alphabets, who (except German speakers) generally expect to see spaces between words too. Japanese writing doesn't use spaces either but all the standard Romaji transliterations do make use of spaces.
