Below I would like to explain some methods how Google can determine the trustworthiness, credibility and authority of a source or entity.
Sentiment analysis as a method for trust assessment
Sentiment analysis is used to evaluate the mood of a sentence or text section. As can be deduced from Google's NLP API, Google is able to carry out sentiment analysis. Sentiment analysis and entity analysis are central sub-steps in natural language processing . In this way, Google can analyze the mood surrounding the mention of entities, including companies and people. If the entity is mentioned frequently in a negative context, it does not seem trustworthy and vice versa. Reviews and ratings in particular could be an interesting source for such analyses.
Example of a sentiment analysis via the Google NLP API, source: digitale-wunderwelt.de
Here is an excerpt from Google’s NLP API documentation :
Per-entity sentiment analysis combines entity analysis with sentiment analysis and attempts to identify the attitude (positive or negative) expressed in the entities of the text. Each entity sentiment is represented by numerical score and magnitude values and is determined for each mention of an entity. These scores are then summarized into an overall sentiment score (score and magnitude) for an entity…
The Natural Language API processes the given text to extract the entities and determine the sentiment. A request for a per-entity sentiment analysis returns a response that contains the following information: the s found in the document content entities, the entry mentions for each mention of the entity, and the numerical values of score and magnitude for each mention, as described in Interpreting Sentiment Analysis Values . The scoreand magnitudevalues for an entity are a summary of the specific scoreand magnitudevalues for each mention of the entity. The scoreand magnitudevalues for an entity can 0 be when the sentiment in the text yields a low score, magnitude resulting in a score of 0, or when the sentiment is mixed, resulting in a score score of 0.
See also the following patent.
Sentiment detection as a ranking signal for reviewable entities
This Google patent was signed in the latest version in October 2017. The patent describes in part the process of sentiment analysis via Google's NLP API.
The patent describes how sentiment analysis can be used to identify sentiments around assessable entities in documents. The results can then be used to rank entities and associated documents. Assessable paraguay phone number data entities include people, places or things about which an opinion can be expressed, such as restaurants, hotels, consumer goods such as electronics, films, books and live performances.
Structured, unstructured data and text can be used as sources for sentiment analysis. Structured reviews are collected from popular review websites such as Google Maps, TripAdvisor, Citysearch or Yelp. Structured reviews can also be collected from other types of text documents such as the text of books, newspapers and magazines.
Unstructured reviews are text documents that reference the verifiable entity and are likely to contain an opinion about the verifiable entity. Unstructured reviews contain a text review but no rating such as star ratings. Unstructured reviews typically contain sentiment in documents with less structured formats such as newsgroups or blogs.
The storage and ranking process is carried out using a ranking analysis engine and an entity ranking data repository. The entity ranking data repository consists of a database for sentiments, entity ratings and user interaction. The user interaction database records user behavior in the SERPs with entity-relevant documents. The process includes typical NLP sub-steps such as part of speech tagging.