Sometimes, it’s best to turn off autodetect of the character encoding, and then manually set the correct encoding through the Encoding > Character Sets menu. There is no 100%-accurate encoding-detection method, because those “encodings” are all 8bit encodings, and so are technically using the same 8bit values (bytes) to indicate different characters; there are some hints that will emphasize certain bytes happen with greater or lesser frequency, which is how encoding auto-detect works; but there is no way to be 100% sure of which of those encodings are used without embedded metadata.
(Or always edit files that use a full-unicode-compatible encoding, such as UTF8 or UTF-16/UCS2 – but of course, your data isn’t in that format; though you could use iconv or similar tool; if you search the forum for “iconv”, you will see some example usages of it)