Should flag "n times than"

I just heard this for the first time in a YouTube video, but examples are not hard to find:

- The digit in tens place of two digit number is **three times than** that in the units place.
- A note book costs **four times than** a pencil.
- The age of the mother is **five times than** of her daughter.
- Then, of the resulting numbers, the larger one is **five times than** the smaller one.
- Usually it is **two times than** realtime unless your hardware is slow.
- Hudi total upsert time is **twice than** the individual jobs time in Spark UI added together
- High cpu usage when idle (**twice than** 1.18)
- [TFRecords File is too large] **3 times than** parquet, **1.3 times than** normal textFile
- Framerate is always **twice than** the monitor refresh
- Complex Calculate use avx2 are slower **3 times than** normal
- result length of encrypt here is **twice than** which in Java, but backend can decrypt it correctly
- it has the wrong bbox (like **2 times than** real) that already overlaps with each character
- no response_model is about **4 times than** having response_model in efficiency.
- inference time spent about **2 times than** trtmodel
- In our case, the duration of docker push in worst case is **ten times than** before, and usually about **five times than** before.

There are also quite a few false positives so a naïve implementation wouldn't work.

No file or test is named in the issue. Start by locating the grammar rule definitions and their tests, then use the listed “n times than” examples and false-positive warning to define coverage; done means valid instances are flagged without a naïve match.

Written by the indexing model from the issue text.