Natural language processing techniques using joint sentiment-topic modeling

An apparatus for generating a predicted document priority score for a digital document, the apparatus comprising at least one processor and at least one memory including program code, the at least one memory and the program code configured to, with the processor, cause the apparatus to at least:

• identify a joint sentiment-topic (JST) latent distribution of the digital document, wherein the JST latent distribution describes one or more topic designation probabilities for the digital document and one or more sentiment designation probabilities for the digital document;
• determine, by processing the one or more topic designation probabilities using a document-topic entropy determination machine learning model, a document-topic entropy measure for the digital document;
• determine, by processing the one or more sentiment designation probabilities using a document-topic entropy determination machine learning model, a sentiment-topic entropy measure for the digital document;
• determine, by processing one or more per-word inverse domain frequency measures for the digital document using a document popularity determination machine learning model, a document popularity measure for the digital document;
• generate the predicted document priority score based at least in part on the document-topic entropy measure, the sentiment-topic entropy measure, and the document popularity measure;
• and
• perform one or more prediction-based actions based at least in part on the predicted document priority score.

A computer program product for generating a predicted priority score for a digital document, the computer program product comprising at least one non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions configured to:

• identify a joint sentiment-topic (JST) latent distribution of the digital document, wherein the JST latent distribution describes one or more topic designation probabilities for the digital document and one or more sentiment designation probabilities for the digital document;
• determine, by processing the one or more topic designation probabilities using a document-topic entropy determination machine learning model, a document-topic entropy measure for the digital document;
• determine, by processing the one or more sentiment designation probabilities using a document-topic entropy determination machine learning model, a sentiment-topic entropy measure for the digital document;
• determine, by processing one or more per-word inverse domain frequency measures for the digital document using a document popularity determination machine learning model, a document popularity measure for the digital document;
• generate the predicted document priority score based at least in part on the document-topic entropy measure, the sentiment-topic entropy measure, and the document popularity measure;
• and
• perform one or more prediction-based actions based at least in part on the predicted document priority score.