System and methods for automatically detecting deceptive content
Systems and methods for detecting deceptive opinion spam. Certain embodiments include a classifier with improved accuracy for detecting deceptive opinion entries. A feature analysis of learned models reveals a relationship between deceptive opinions and imaginative writing. By modeling deception in a generative framework, the prevalence of deception in two popular online review communities may be determined. Deceptive opinion spam is a rapidly growing and widespread problem, especially in review communities with minimal posting requirements.
1. A method for classifying textual opinion information as truthful or deceptive, comprising the steps of:
communicating with a processor a first source of opinion information via a communication interface connecting a user system to a network to collect opinion information, wherein the opinion information consists of at least one set of known deceptive opinion information and at least one set of known truthful opinion information forming an initial dataset, the communication interface being a wired communication interface or a wireless communication interface;
storing by the processor the opinion information in a main memory of the user system;
analyzing separately by the processor each of the set of known deceptive opinion information and the set of known truthful opinion information of the opinion information to determine features associated with each set in the initial dataset, wherein the machine-analysis comprises a genre identification approach that reviews each part of speech of the opinion information;
automatically generating a model based on the analyzing step in which a first set of features comprising nouns, adjectives, prepositions, determiners, coordinating conjunctions are associated with the set of known deceptive opinion information and a second set of features comprising verbs, adverbs, pronouns and pre-determiners are associated with the set of known truthful opinion information;
receiving by the processor an online review of a product or a service;
applying by the processor the model to the online review, wherein the processor identifies text of the online review as one or more nouns, adjectives, prepositions, determiners, coordinating conjunctions, verbs, adverbs, pronouns and pre-determiners;
calculating by the processor a first number of features of the first set in the text and a second number of features of the second set in the text; and
categorizing by the processor the online review as deceptive when the first number is greater than the second number and categorizing the online review as truthful when the second number is greater than the first number.
2. The method of claim 1 further comprising the step of:
displaying on a display device a label of one or more online reviews.
3. The method of claim 2 , wherein the label is the word “deceptive”.
4. The method of claim 2 , wherein the label is the word “truthful”.
5. The method of claim 3 further comprising the step of:
removing by the processor the one or more online reviews labeled as “deceptive”.
6. The method of claim 1 further comprising:
displaying on a display device statistics about prevalence of deceptive opinion information.