RTEInformationRetrievalProject

The file called RTEProject contains the first approach used for the development of the project in which 4 methodologies were used to calculate the similarity between each pair of sentences. These methodologies involve the computation of similarity through the use of a library such as spaCy, the Word2vec gensim model, the use of wordnet and the computation of the Levenshtein distance. Instead, the RTEProject1 file contains the second approach in which the sentence-transformers model was used to calculate the similarity of each pair of sentences. The idea of using the two approaches is to be able to compare the results produced by the supervised classification models used within the project. As concern the execution on sentence pairs contained in the MNLI dataset, it is necessary to download the latter and then load the file multinli_1.0_train.jsonl for its use. The dataset is available at this link: https://deepai.org/dataset/multinli