Methods and systems of utilizing machine learning to provide trust scores in an online automobile marketplace

1. A computerized method of providing health scores in an online marketplace comprising:

providing, with at least one computer processor, an item listing, wherein the item listing is listed in an e-commerce marketplace;

identifying the item listing into a set of listing parameters, wherein each listing parameter comprises one or more listing images and one or more terms that are descriptive of the item listing, wherein the listing parameter is based on an image of an automobile uploaded by a user;

based on historical sales data, calculating a set of weights assignments of the set of listing parameters, with a Multivariate Regression model by:

splitting the historical sales data into a training set and a test set,

generating the weights assignments using a multivariate linear regression process, and

applying the weights assignments to the listing parameters; and

using the listing parameters to:

determine a trust score for the item listing, wherein the trust score is based on a number of images of an item in the item listing, a quality score of the number of images, and a description score of a description of the item, wherein the trust score is determined using a specified machine learning algorithm comprising at least one artificial neural network, and, wherein the trust score is further based on a set of basic facts about the item, a set of key factors of the item, and a list of available options for purchasing the item;

determine a pricing score, wherein the pricing score is based on a percentage variation from a geographically relevant item valuation and an item research service, and a condition of the item, and wherein the pricing score is based on a listing price relative to a market price of a similar automobile and an algorithmic pricing engine, and wherein the pricing score is further based on a percentage variation from median price of a set of items that are in a same class as the item and a percentage of other current item listings of the set of items that are below a quoted price in the item listing;

determine a seller score, wherein the seller score is based on the verified status of the seller, a seller rating of the seller, and a percentage of positive feedback, and wherein every seller is unique, hence this score is based on seller related factors comprising the Ratings given to the seller by buyers, whether the seller is verified or not, and, wherein the seller score is further based on a seller engagement score, a showroom score, and a dealership score; and

determine a health score, wherein the health score is based on a seller declaration, a service log, a verification that the item is insured, and a verification that the item is properly registered, and wherein the health score is based on factors which enhance trust on the automobile including an inspection report and a warranty,

wherein the item comprises a used automobile, and

wherein the pricing score is further based on a percentage variation from median price of a set of items that are in a same class as the item and a percentage of other current item listings of the set of items that are below a quoted price in the item listing.

2. The computerized method of claim 1 further comprising:

generating a full-health score based on a weighted average of the trust score, the pricing score, the seller score and the health score.

3. A server system comprising:

a processor configured to execute instructions;

a memory containing instructions which, when executed on the processor, cause the processor to perform operations that:

provide, with at least one computer processor, an item listing, wherein the item listing is listed in an e-commerce marketplace;

identify the item listing into a set of listing parameters, wherein each listing parameter comprises one or more listing images and one or more terms that are descriptive of the item listing, wherein the listing parameter is based on an image of the automobile uploaded by a user;

based on historical sales data, calculate a set of weights assignments of the set of listing parameters values with a Multivariate Regression model by:

splitting the historical sales data into a training set and a test set,

generating the weights assignments using a multivariate linear regression process, and

applying the weights assignments to the listing parameters; and

using the listing parameters to:

determine a trust score for the item listing, wherein the trust score is based on a number of images of an item in the item listing, a quality score of the number of images, and a description score of a description of the item, wherein the trust score is determined using a specified machine learning algorithm comprising at least one artificial neural network, and wherein the trust score is further based on a set of basic facts about the item, a set of key factors of the item, and a list of available options for purchasing the item;

determine a pricing score, wherein the pricing score is based on a percentage variation from a geographically relevant item valuation and an item research service, and a condition of the item, and wherein the pricing score is based on a listing price relative to a market price of a similar automobile and an algorithmic pricing engine, and wherein the pricing score is further based on a percentage variation from median price of a set of items that are in a same class as the item and a percentage of other current item listings of the set of items that are below a quoted price in the item listing;

determine a seller score, wherein the seller score is based on the verified status of the seller, a seller rating of the seller, and a percentage of positive feedback, and wherein every seller is unique, hence this score is based on seller related factors comprising the Ratings given to the seller by buyers, whether the seller is verified or not, and wherein the seller score is further based on a seller engagement score, a showroom score, and a dealership score; and

determine a health score, wherein the health score is based on a seller declaration, a service log, a verification that the item is insured, and a verification that the item is properly registered, and wherein the health score is based on factors which enhance trust on the automobile including an inspection report and a warranty,

wherein the item comprises a used automobile, and

wherein the pricing score is further based on a percentage variation from median price of a set of items that are in a same class as the item and a percentage of other current item listings of the set of items that are below a quoted price in the item listing.

4. The server system of claim 3 , wherein the trust score is further based on a set of basic facts about the item, a set of key factors of the item, and a list of available options for purchasing the item.

5. The server system of claim 4 , wherein the seller score is further based on a seller engagement score, a showroom score, and a dealership score.