This project models 400 years of semantic shift in Spanish words and trains a classifier to detect whether words underwent change between two time periods. In addition, it examines whether the amount of context used to predict a word’s embedding can beoptimized to improve classification results and analyzes whether the optimal context window sizevaries for words in different lexical categories. The models used draw on new methods of analyzing semantic shift by creating vector embeddings for words trained via neural networks, after which asupport vector classifier is trained to detect semantic shift using a new dataset of words annotated with their historical change.
This code accompanies the paper "Automatic Detection of Semantic Shift in Spanish with Context Optimization".
This project is the work of Eve Fleisig under the advice of Professor Christiane Fellbaum and was created for an independent work paper in Fall 2019.