Systems and methods for modeling item similarity and correlating item information

1. A system for correlating item data, the system comprising:

at least one processor; and

a non-transitory computer-readable medium containing a set of instructions that, when executed by the at least one processor, cause the processor to perform steps comprising:

receiving reference text data associated with a reference item from a remote device;

receiving reference image data associated with the reference item from the remote device;

determining candidate text data and candidate image data associated with at least one candidate item;

selecting a text correlation model;

determining a first similarity score by applying the text correlation model to the reference text data and the candidate text data;

selecting an image correlation model;

determining a second similarity score by applying the image correlation model to the reference image data and the candidate image data;

calculating a confidence score based on the first and second similarity scores; and

performing a responsive action based on the calculated confidence score, wherein the responsive action comprises at least one of: creating an association, changing the text correlation model, or changing the image correlation model.

2. The system of claim 1 , wherein the steps further comprise:

determining whether the confidence score falls below a threshold; and

when the confidence score falls below the threshold, determining a differentiation factor indicating a difference between the reference item and the at least one candidate item; and

wherein the responsive action comprises at least one of: adjusting a parameter of the text correlation model or the image correlation model using the differentiation factor; or adding a new parameter to the text correlation model or image correlation model based on the differentiation factor.

3. The system of claim 2 , wherein the differentiation factor is associated with a difference between a first item specification of the reference item and a second item specification of the candidate item.

4. The system of claim 3 , wherein the first item specification and the second item specification each comprises at least one of: a color, a dimension, a model number, a weight, a shape, a scent, a material, a time of production, a multi-part item, or an item feature.

5. The system of claim 1 , wherein the steps further comprise:

determining whether the confidence score is equal to or greater than a threshold; and

when the confidence score is equal to or greater than the threshold:

creating an association between the reference item and the candidate item;

monitoring a webpage associated with the reference item to detect a change in information associated with the reference item at the webpage; and

transmitting a notification to a user device upon detecting the change in information.

6. The system of claim 5 , wherein the change is associated with a price of the reference item at the monitored webpage.

7. The system of claim 1 , wherein the image correlation model is a random forest model.

8. The system of claim 1 , wherein at least one of the text correlation model or image correlation model is selected based on a website or an entity associated with the remote device.

9. The system of claim 1 , wherein at least one of the text correlation model or image correlation model is selected based on a category of the reference item.

10. The system of claim 1 , wherein determining the candidate text data and candidate image data comprises:

tokenizing the reference text data;

comparing the tokenized reference text data to item data stored in a database, the item data comprising pairs of reference item images and reference item text; and

selecting a subset of the item data as the candidate text data and candidate image data based on the comparison.

11. The system of claim 10 , wherein:

the steps further comprise tagging a portion of the reference text data as a price of the reference item; and

selecting the subset of the item data comprises selecting item data including reference item text indicating a price within a predetermined range of the tagged price.

12. The system of claim 1 , wherein:

applying the image correlation model to the reference image data comprises applying at least one of a cropping operation, a re-sizing operation, a brightness alteration operation, a contrast operation alteration, or an interpolation operation to the reference image data; and

the image correlation model is image resolution-agnostic.

13. The system of claim 1 , wherein:

the reference text data is crawled from a webpage and tagged by a first web crawler;

the reference image data is crawled by a second web crawler;

the reference image data comprises multiple images crawled from a single item page of the reference item; and

applying the image correlation model to the reference image data comprises comparing the reference image data to the candidate image data.

14. The system of claim 13 , wherein:

comparing the reference image data to the candidate image data includes performing a plurality of image comparisons;

applying the image correlation model to the reference image data comprises calculating a third similarity score for each of the image comparisons; and

the second similarity score is based on the third similarity scores.

15. The system of claim 13 , wherein:

comparing the reference image data to the candidate image data comprises performing a plurality of image comparisons;

applying the image correlation model to the reference image data comprises calculating a third similarity score for each of the image comparisons; and

the second similarity score is a maximum of the third similarity scores.

16. The system of claim 1 , wherein the text correlation model contains a text frequency parameter having a weight that is inversely related to a frequency of a character combination in a reference dataset.

17. The system of claim 1 , wherein:

the text correlation model is trained to ignore a property of the reference text data when determining the first similarity score; or

the image correlation model is trained to ignore a property of the reference image data when determining the second similarity score.

18. The system of claim 15 , wherein the ignored property is based on a user input.

19. A computer-implemented method for correlating item data comprising:

receiving reference text data associated with a reference item from a remote device;

receiving reference image data associated with the reference item from the remote device;

determining candidate text data and candidate image data associated with at least one candidate item;

selecting a text correlation model;

determining a first similarity score by applying the text correlation model to the reference text data and the candidate text data;

selecting an image correlation model;

determining a second similarity score by applying the image correlation model to the reference image data and the candidate image data;

calculating a confidence score based on the first and second similarity scores; and

performing a responsive action based on the calculated confidence score, wherein the responsive action comprises at least one of: creating an association, changing the text correlation model, or changing the image correlation model.

20. A system for correlating item data, the system comprising:

a relational database storing associations between item data;

a first computing device comprising:

at least one processor; and

a non-transitory computer-readable medium containing a set of instructions that, when executed by the at least one processor, cause the processor to perform steps comprising:

crawling a website to obtain text data associated with a reference item;

tagging at least one data element of the obtained text data as a title or a price;

crawling a website to obtain image data associated with the reference item; and

transmitting the text data associated with the reference item and the image data associated with the reference item to a second computing device; and

the second computing device comprising:

at least one processor; and

a non-transitory computer-readable medium containing a set of instructions that, when executed by the at least one processor, cause the processor to perform steps comprising:

receiving the transmitted text and image data;

determining candidate text data and candidate image data associated with at least one candidate item;

selecting a text correlation model;

determining a first similarity score by applying the text correlation model to the reference text data and the candidate text data;

selecting an image correlation random forest model;

determining a second similarity score by applying the image correlation random forest model to the reference image data and the candidate image data;

calculating a confidence score based on the first and second similarity scores; and

based on the calculated confidence score, performing at least one of: modifying a parameter of the image correlation random forest model, adding a new parameter to the image correlation random forest model, modifying an association stored at the relational database, or adding a new association to the relational database.