سلام دوستان
این یه تیکه از یه پایان نامس ترجش میکنم ولی ترجمه ی کلیش خیلی پرت میشه اصلا با نتایج همخوانی نداره احساس میکنم اشتباه دارم ترجمه میکنم
میشه کمکم کنید :
این متن اصلی
The LTFIDF metrics was also another poor performer, although it received higher overlap
scores than the three mentioned above. This result is slightly puzzling and may need further
investigation. In term weighting methods for information retrieval the logged version of IDF is
used more frequently than the non-logged version, yet here we see poorer results. A possible
explanation for this is the information retrieval typically deals with very large corpora where
logging document numbers is more drastic. Corpus term frequency and TFIDF did relatively
well though, so lessening the impact of IDF should make the results look more like TF, and
clearly it does not. An explanation for this could also be that because the corpora were so small
that logging most corpus sizes produced small decimal values which greatly dampened term
frequency. This dampening effect caused the term scores to be more closely lumped, which may
in turn produce more errors.
