search

KshitijKarthick / tvecs

Establish Semantic Relatedness across Languages Documentation - http://kshitijkarthick.github.io/tvecs
https://tvecs.kshitijkarthick.me
MIT License
3 stars 0 forks source link

Obtain Correlation Score b/w Human Scores & T-Vecs Model #7

Closed KshitijKarthick closed 8 years ago

KshitijKarthick commented 8 years ago

Generate Correlation Score Between Human Score & T-Vecs Model generated for semantically similar words.

KshitijKarthick commented 8 years ago

Have a look at the human semantic similarity score after utilising Yandex for translation service with the cosine distance obtained from T-Vecs Model. https://drive.google.com/open?id=0B08EScGNCpQpYTN2MmxkRXV5djA

Prateeksha13 commented 8 years ago

New module "evaluation" is created under t-vecs -> modules. Correlation score is computed between Human Semantic Similarity Score and T-Vecs Model using SciPy statistics function pearsonr which computes Pearson's correlation coefficient and p-value. Correlation coefficient - Measure of degree of relatedness between two variables. p-value - The null hypothesis is that the two variables are uncorrelated. The p-value is a number between zero and one that represents the probability that your data would have arisen if the null hypothesis were true.

Commit id : a37932a