Methods and systems for assessing excessive accessory listings in search results

Abstract

A system and method for generating a score for a listing are described. A data field of a listing is parsed and at least one element from the data field is generated. A score for the listing is calculated based on the at least one element, the listing score representing a probability of the listing being in one of two binary classifications. The listing score and one or more listing attribute values are inputted into a binary classifier. An output is generated using the binary classifier, the output representing a refined score for the listing based on the listing score and at least one of the listing attribute values.

1 . (canceled) 
     
     
         2 . A system, comprising:
 a processor-implemented textual mining module configured to parse a data field of a listing and generate at least one element from the data field;   a processor-implemented scoring module configured to:
 calculate a score for the listing based on the at least one element, the listing score representing a probability of the listing being in one of two binary classifications; and 
   a processor-implemented binary classifier module configured to generate an output representing a refined score for the listing based on the listing score and at least one listing attribute value.   
     
     
         3 . The system of  claim 2 , wherein the processor-implemented scoring module uses a naive Bayes classifier to calculate a score of the at least one element. 
     
     
         4 . The system of  claim 2 , wherein the listing is an item listing of an item offered for sale, and wherein the two binary classifications are products and product accessories. 
     
     
         5 . The system of  claim 2 , wherein the processor-implemented scoring module is configured to calculate the element score by determining a first number of occurrences of the at least one element in a first binary classification in a set of item listings and a second number of occurrences of the at least one element in a second binary classification in the set of item listings and by obtaining a ratio of the first number of occurrences to a sum of the first number of occurrences and the second number of occurrences. 
     
     
         6 . The system of  claim 5 , wherein the processor-implemented scoring module is further configured to calculate the element score by normalizing the ratio to derive the element score. 
     
     
         7 . The system of  claim 5 , wherein the processor-implemented scoring module is configured to calculate the listing score by adding together the first number of occurrences for a plurality of the at least one element and dividing the added first number of occurrences for the at least one element by the sum of the first number of occurrences and the second number of occurrences. 
     
     
         8 . The system of  claim 4 , wherein the at least one listing attribute value includes at least one of quantity of the item, price of the item, median price of the item, product sales rank, shipping cost, return policy, seller feedback score, and item location. 
     
     
         9 . A computer-implemented method, comprising:
 parsing, by at least one processor, a data field of a listing and generating at least one element from the data field;   calculating, by the at least one processor, a score for the listing based on the at least one element, the listing score representing a probability of the listing being in one of two binary classifications;   inputting the listing score and one or more listing attribute values into a binary classifier; and   generating an output using the binary classifier, the output representing a refined score for the listing based on the listing score and at least one of the listing attribute values.   
     
     
         10 . The computer-implemented method of  claim 9 , wherein the listing is an item listing of an item offered for sale, and wherein the two binary classifications are products and product accessories. 
     
     
         11 . The computer-implemented method of  claim 9 , wherein a calculating of at least one element score uses a naïve Bayes classifier and comprises:
 determining a first number of occurrences of the at least one element in a first binary classification in a set of item listings and a second number of occurrences of the at least one element in a second binary classification in the set of item listings; and 
 calculating a ratio of the first number of occurrences to a sum of the first number of occurrences and the second number of occurrences. 
 
     
     
         12 . The computer-implemented method of  claim 11 , wherein the calculating of the at least one element score further comprises normalizing the ratio to derive the at least one element score. 
     
     
         13 . The computer-implemented method of  claim 11 , wherein the calculating of the listing score comprises:
 adding together the first number of occurrences for a plurality of the at least one element; and   dividing the added first number of occurrences for the at least one element by the sum of the first number of occurrences and the second number of occurrences.   
     
     
         14 . The computer-implemented method of  claim 9 , wherein the listing attribute values include at least one of quantity of the item, price of the item, median price of the item, product sales rank, shipping cost, return policy, seller feedback score, and item location. 
     
     
         15 . A non-transitory machine-readable storage medium storing a set of instructions that, when executed by at least one processor, causes the at least one processor to perform operations comprising:
 parsing a data field of a listing and generating at least one element from the data field;   calculating a score for the listing based on the at least one element, the listing score representing a probability of the listing being in one of two binary classifications;   inputting the listing score and one or more listing attribute values into a binary classifier; and   generating an output using the binary classifier, the output representing a refined score for the listing based on the listing score and at least one of the listing attribute values.   
     
     
         16 . The machine-readable storage medium of  claim 15 , wherein the listing is an item listing of an item offered for sale, and wherein the two binary classifications are products and product accessories. 
     
     
         17 . The machine-readable storage medium of  claim 15 , wherein the calculating of the at least one element score comprises:
 determining a first number of occurrences of the at least one element in a first binary classification in a set of item listings and a second number of occurrences of the at least one element in a second binary classification in the set of item listings; and   calculating a ratio of the first number of occurrences to a sum of the first number of occurrences and the second number of occurrences.   
     
     
         18 . The machine-readable storage medium of  claim 17 , wherein the calculating of the at least one element score further comprises normalizing the ratio to derive the at least one element score. 
     
     
         19 . The machine-readable storage medium of  claim 17 , wherein a calculating of at least one element score uses a naïve Bayes classifier and comprises:
 adding together the first number of occurrences for a plurality of the at least one element; and 
 dividing the added first number of occurrences for the at least one element by the sum of the first number of occurrences and the second number of occurrences. 
 
     
     
         20 . The machine-readable storage medium of  claim 15 , wherein the listing attribute values include at least one of quantity of the item, price of the item, median price of the item, product sales rank, shipping cost, return policy, seller feedback score, and item location.