Sentiment polarity for users of a social networking system

A method comprising:

• identifying content of a page of a social networking system;
• receiving information about a first interaction between a user of the social networking system and the page;
• inferring by a computer system a sentiment polarity of the user based on the received information about the first interaction, the sentiment polarity indicating a positive or negative sentiment of the user toward the content of the page by:
• identifying a set of trusted users associated with the content, the set of trusted users comprising users of the social networking system possessing a strong sentiment polarity toward the content
• determining, for each trusted user, a data set comprising information about interactions between the trusted user and the page in conjunction with the trusted user's sentiment polarity;
• training a machine learning model using the determined data sets as a training set and
• inferring the sentiment polarity of the user using the trained machine learning model based on the received information;
• storing the inferred sentiment polarity of the user toward the content of the page;
• receiving a second interaction from the user related to the content;
• and
• associating with the second interaction the stored sentiment polarity of the user toward the content of the page.

A method comprising:

• identifying content of a page of a social networking system;
• receiving information about a first interaction between a non-trusted user of the social networking system and the page;
• inferring by a computer system a sentiment polarity of the non-trusted user based on the received information about the first interaction, the sentiment polarity indicating a first sentiment of the non-trusted user toward the content by:
• identifying a set of trusted users associated with the content, the set of trusted users comprising users of the social networking system possessing a strong sentiment polarity toward the content
• determining, for each trusted user, a data set comprising information about interactions between the trusted user and the page in conjunction with the trusted user's sentiment polarity;
• training a machine learning model using the determined data sets as a training set and
• inferring the sentiment polarity of the user using the trained machine learning model based on the received information;
• receiving a second interaction from a trusted user related to the content indicating a second sentiment toward the first interaction by the non-trusted user;
• and
• adjusting the first sentiment of the non-trusted user toward the content in a direction corresponding to the second sentiment indicated by the trusted user responsive to determining that the first sentiment is different from the second sentiment.

A computer program product comprising a non-transitory computer-readable storage medium containing computer program code for:

• identifying text content of a page of a social networking system;
• extracting a topic from the text content;
• receiving information about a first interaction between a user of the social networking system and the page;
• determining a sentiment of the user toward the text content based on the received information about the first interaction;
• inferring by a computer system that the user has a sentiment polarity toward the extracted topic based on the received information about the first interaction, the sentiment polarity corresponding to the user's sentiment toward the text content and indicating a positive or negative sentiment of the user toward the topic by:
• identifying a set of trusted users associated with the text content, the set of trusted users comprising users of the social networking system possessing a strong sentiment polarity toward the text content
• determining, for each trusted user, a data set comprising information about interactions between the trusted user and the page in conjunction with the trusted user's sentiment polarity;
• training a machine learning model using the determined data sets as a training set and
• inferring the sentiment polarity of the user using the trained machine learning model based on the received information;
• storing the inferred sentiment polarity of the user toward the extracted topic;
• receiving a second interaction from the user related to the extracted topic;
• and
• associating with the second interaction the stored sentiment polarity of the user toward the extracted topic.

A computer program product comprising a non-transitory computer-readable storage medium containing computer program code for:

• identifying content of a page of a social networking system;
• receiving information about a first interaction between a user of the social networking system and the page;
• inferring by a computer system a sentiment polarity of the user based on the received information about the first interaction, the sentiment polarity indicating a positive or negative sentiment of the user toward the content of the page by:
• identifying a set of trusted users associated with the content, the set of trusted users comprising users of the social networking system possessing a strong sentiment polarity toward the content
• determining, for each trusted user, a data set comprising information about interactions between the trusted user and the page in conjunction with the trusted user's sentiment polarity;
• training a machine learning model using the determined data sets as a training set and
• inferring the sentiment polarity of the user using the trained machine learning model based on the received information;
• storing the inferred sentiment polarity of the user toward the content of the page;
• receiving a second interaction from the user related to the content;
• and
• associating with the second interaction the stored sentiment polarity of the user toward the content of the page.