Fingerprint-based literary works recommendation system

Abstract

A system that recommends literary works to a user based on identified trends of how text in the literary works liked and/or disliked by the user are written and/or structured is provided. For example, the system may analyze the text of a literary work to identify one or more metrics. Based on the identified metrics, the system can generate an analytical summary called a fingerprint for the literary work. The ratings assigned to literary works by the user may be used in conjunction with the generated fingerprints to generate positive and/or negative models for the user. The positive model captures aspects of literary works that the user likes and the negative model captures aspects of literary works that the user dislikes. The system can then compare some or all of the generated fingerprints in a literary works fingerprint database with the positive and/or negative models to select literary works to recommend to the user.

What is claimed is: 
     
         1 . A system comprising:
 a literary works database comprising text of a plurality of literary works; and   a server system comprising one or more computing devices, the server system comprising executable code that, when executed, causes the server system to:
 for one or more literary works in the plurality of literary works,
 access text of the respective literary work from the literary works database, 
 compute a set of metrics comprising semantic metrics and syntactic metrics, wherein the computation of the set of metrics comprises a parsing of the accessed text, 
 generate a fingerprint based on the computed set of metrics, and 
 store the generated fingerprint in association with the respective literary work; 
 
 receive an indication to provide a recommendation to a user device associated with a first user, wherein one or more ratings are associated with the first user, and wherein each rating in the one or more ratings associated with the first user is assigned to a literary work in the plurality of literary works by the first user; 
 retrieve a model associated with the first user that indicates literary works in the plurality of literary works favorably rated by the first user, wherein the model is generated based on a first set of literary works in the plurality of literary works that are assigned a rating above a first threshold value by the first user; 
 identify one or more second literary works in the plurality of literary works that are associated with a fingerprint stored in the literary works fingerprint database that has at least one metric that is within a second threshold value of at least one metric of the model; and 
 transmit an identifier of the one or more second literary works to the user device. 
   
     
     
         2 . The system of  claim 1 , further comprising a user database comprising an identification of a plurality of users and one or more ratings associated with at least some users, wherein each rating is assigned to a literary work by the respective user. 
     
     
         3 . The system of  claim 1 , wherein the text of the respective literary work is in an electronic format, and wherein the accessed text is parsed using natural language processing 
     
     
         4 . The system of  claim 1 , wherein the executable code, when executed, further causes the server system to:
 retrieve a second model associated with the first user, wherein the second model is generated based on a second set of literary works in the plurality of literary works that are assigned a rating below a third threshold value by the first user;   identify one or more third literary works in the plurality of literary works that are associated with a fingerprint that is within a fourth threshold value of the second model; and   prevent transmission of an identity of the one or more third literary works to the user device.   
     
     
         5 . A computer-implemented method of analyzing a corpus of text to provide a recommendation to a user, the method comprising:
 as implemented by a computer system comprising one or more computing devices, the computer system configured with specific executable instructions,   receiving an indication to provide a recommendation to a user device associated with a first user, wherein one or more ratings are associated with the first user, and wherein each rating is assigned to a literary work by the first user;   retrieving a model associated with the first user, wherein the model is generated based on a first set of literary works that are assigned a rating above a first threshold value by the first user;   identifying one or more second literary works in the plurality of literary works that are associated with a fingerprint that is within a second threshold value of the model; and   transmitting an identifier of the one or more second literary works to the user device.   
     
     
         6 . The computer-implemented method of  claim 5 , further comprising:
 retrieving a second model associated with the first user, wherein the second model is generated based on a second set of literary works that are assigned a rating below a third threshold value by the first user;   identifying one or more third literary works in the plurality of literary works that are associated with a fingerprint that is within a fourth threshold value of the second model; and   preventing transmission of an identifier of the one or more third literary works to the user device.   
     
     
         7 . The computer-implemented method of  claim 5 , further comprising:
 receiving an indication that the first set of literary works includes a new literary work;   determining an updated model based on fingerprints associated with the first set of literary works.   
     
     
         8 . The computer-implemented method of  claim 7 , further comprising:
 receiving an indication to provide a second recommendation to the user device associated with the first user;   identifying one or more third literary works in the plurality of literary works that are associated with a fingerprint that is within the second threshold value of the updated model; and   transmitting an identifier of the one or more third literary works to the user device.   
     
     
         9 . The computer-implemented method of  claim 5 , further comprising retrieving at least one fingerprint stored in a literary works fingerprint database. 
     
     
         10 . The computer-implemented method of  claim 9 , wherein the retrieved fingerprints are generated based on semantic metrics and syntactic metrics. 
     
     
         11 . The computer-implemented method of  claim 5 , further comprising comparing the model with retrieved fingerprints. 
     
     
         12 . The computer-implemented method of  claim 11 , wherein comparing the model with the retrieved fingerprints comprises:
 selecting metrics of the model and metrics of the retrieved fingerprints using machine-learning techniques; and   comparing the selected metrics of the model with the selected metrics of the retrieved fingerprints.   
     
     
         13 . The computer-implemented method of  claim 5 , further comprising:
 ranking the one or more second literary works according to a proximity of a fingerprint associated with a respective second literary work to the model; and   transmitting the ranking of the one or more second literary works to the user device.   
     
     
         14 . A system for analyzing a corpus of text to generate a fingerprint that can be used to provide a recommendation to a user, comprising:
 a literary works database comprising text of a plurality of literary works;   a literary works fingerprint database configured to store fingerprints for literary works in the plurality; and   a text analyzer configured to:
 access text of a first literary work in the plurality of literary works from the literary works database; 
 compute a set of metrics comprising semantic metrics and syntactic metrics, wherein the computation of the set of metrics comprises a parsing of the accessed text; 
 generate a fingerprint based on the computed set of metrics; and 
 store the generated fingerprint in association with the first literary work in the literary works fingerprint database for comparison with an aggregate of fingerprints of other literary works assigned a rating by a user above a threshold value to identify whether the first literary work is similar to the other literary works. 
   
     
     
         15 . The system of  claim 14 , wherein the generated fingerprint comprises a set of values for each metric in the computed set of metrics. 
     
     
         16 . The system of  claim 14 , wherein the text of the first literary work is in an electronic format, and wherein the text analyzer parses the accessed text using natural language processing. 
     
     
         17 . The system of  claim 14 , further comprising a recommendation module configured to transmit an identifier of the first literary work to a computing device associated with the user in connection with a determination that the stored fingerprint is within a second threshold value of the aggregate of fingerprints. 
     
     
         18 . The system of  claim 14 , wherein the text analyzer is further configured to parse the accessed text and retrieve external data to compute the set of metrics. 
     
     
         19 . The system of  claim 18 , wherein the external data comprises values for at least one of word or lexical complexity or density. 
     
     
         20 . The system of  claim 14 , wherein the text analyzer is further configured to receive a notification that the text of the first literary work is available.