In our latest medium.com blog post, our Data Scientist Jan Paulus uses code examples to illustrate how to use the machine learning concept of sentence embeddings and what kind of weighting strategies can be used.
Finding significant features as input for machine learning models is a fundamental element of the Data Team's work at snapADDY. One of the most successful approaches is the well-known word embedding, which allows assigning similar values to words with similar meanings. The same function is sought in many use cases, not only with single words but with entire sentences. The new article on the snapADDY technical blog discusses how to calculate sentence embeddings based on previously trained word embeddings and what kind of strategies can be used for aggregation.