machine learning - Semantic Relations finding algorithms -
my final year project based on machine learning , natural language processing. implementing search bar write query of 2-3 words. after classification of text, want find out semantic relations between 2 words.
for semantic stage -
input - matrix of (rxc) rows words , columns contexts or paragraphs. , each cell have count of finding words contexts.
output - yes or no. particular document. whether there in document.
your answer not clear can give details main approaches in relation extraction (re) task:
- supervised re has employed kernel-based approaches : see link
- distant supervision introduced in bioinformatics : main idea of approach create own training data heuristically matching contents of relation repositories corresponding text (see link)
- unsupervised re collect co-occurrences of word pairs strings between them, , calculate term co-occurrence or generate surface patterns (see link)
we can better if give more details labeled data , real goal of project.
Comments
Post a Comment