Routine item recommendations

1. A system comprising:

a plurality of computing devices, each device of the plurality of computing devices being associated with a customer of a plurality of customers associated with a website;

a server configured to host the website; and

a memory having instructions stored thereon, and a processor configured to read the instructions to:

for each customer of the plurality of customers:

receive, from a corresponding computing device of the plurality of computing devices, anchor item data of an anchor item, the anchor item data including descriptive information metadata;

determine categorical data of a category of a plurality of categories that corresponds to the anchor item;

identify a routine of a plurality of predetermined routines that corresponds to the categorical data of the anchor item, each routine of the plurality of predetermined routines being (i) based on an item purchasing behavior of the corresponding customer of the plurality of customers associated with the website and (ii) associated with a set of categorical data indicating a group of categories that are associated with the corresponding routine;

generate relevant categorical data within the identified routine by applying a ranking algorithm to the categorical data of the identified routine and the categorical data of the anchor item, the relevant categorical data indicating one or more categories that are associated with the identified routine and are relevant to the anchor item, the one or more categories being associated with one or more relevant items;

generate relevant item data from the relevant categorical data by applying an item recommendation model to item data that corresponds to the relevant categorical data;

based on the relevant item data, generate, for each of the one or more relevant items, a graphical representation and a selectable feature, the selectable feature, when triggered, implements a set of operations that add the corresponding relevant item of the one or more relevant items to an online shopping cart associated with the website and the corresponding customer, along with the anchor item.

2. The system of claim 1 , wherein the anchor item data corresponds to an item placed in an online shopping cart.

3. The system of claim 1 , wherein identifying the routine of the plurality of predetermined routines includes:

applying a word embedding model to purchase history data and descriptive information data of items purchased by a customer to generate sets of embeddings; and

grouping the sets of embeddings together to generate categories of items.

4. The system of claim 3 , wherein the processor is further configured to read the instructions to store the determined routine as a predetermined routine.

5. The system of claim 1 , wherein the processor is further configured to read the instructions to apply the ranking algorithm to generate a category correlation score for each category of the plurality of categories, and rank the plurality of categories based on the category correlation score for each category of the plurality of categories.

6. The system of claim 1 , wherein the item recommendation model comprises at least one of an item-to-item co-purchase model and a category trending model.

7. The system of claim 1 , wherein the processor is further configured to read the instructions to apply the item recommendation model to output complementary scores for items corresponding to the item data that corresponds to the relevant categorical data, and to rank the items within the one or more categories of the relevant categorical data based on relevancy to the anchor item that corresponds to the anchor item data.

8. A computer-implemented method comprising:

for each customer of a plurality of customers associated with a website:

receiving, by a processor and from a corresponding computing device of the customer, anchor item data of an anchor item, the anchor item data including descriptive information metadata;

determine, by the processor, categorical data of a category of a plurality of categories that corresponds to the anchor item;

identifying, by the processor, a routine of a plurality of predetermined routines that corresponds to the categorical data of the anchor item, each routine of the plurality of predetermined routines being (i) based on an item purchasing behavior of the corresponding customer of the plurality of customers associated with the website and (ii) associated with a set of categorical data indicating a group of categories that are associated with the corresponding routine;

generating, by the processor, relevant categorical data within the identified routine by applying a ranking algorithm to the categorical data of the identified routine and the categorical data of the anchor item, the relevant categorical data indicating one or more categories that are associated with the identified routine and are relevant to the anchor item, the one or more categories being associated with one or more relevant items;

generating, by the processor, relevant item data from the relevant categorical data by applying an item recommendation model to item data that corresponds to the relevant categorical data; and

based on the relevant item data, generating, by the processor and for each of the one or more relevant items, a graphical representation and a selectable feature, the selectable feature, when triggered, implements a set of operations that add the corresponding relevant item of the one or more relevant items to an online shopping cart associated with the website and the corresponding customer, along with the anchor item.

9. The method of claim 8 , wherein the anchor item data corresponds to an item placed in an online shopping cart.

10. The computer-implemented method of claim 8 , wherein identifying the routine of the plurality of predetermined routines includes:

applying a word embedding model to purchase history data and descriptive information data of items purchased by a customer to generate sets of embeddings; and

grouping the sets of embeddings together to generate categories of items.

11. The computer-implemented method of claim 10 , further comprising storing the determined routine as a predetermined routine.

12. The computer-implemented method of claim 8 , further comprising:

applying the ranking algorithm to generate a category correlation score for each category of the plurality of categories; and

ranking the plurality of categories based on the category correlation score for each category of the plurality of categories.

13. The computer-implemented method of claim 8 , further comprising:

applying the item recommendation model to output complementary scores for items corresponding to the item data; and

ranking the items within a category based on relevancy to the anchor item that corresponds to the anchor item data.

14. A computer program product comprising:

a non-transitory computer readable medium having program instructions stored thereon, the program instructions executable by one or more processors, the program instructions comprising:

for each customer of a plurality of customers associated with a website:

receiving, from a corresponding computing device of the customer, anchor item data of an anchor item, the anchor item data including descriptive information metadata;

determining categorical data of a category of a plurality of categories that corresponds to the anchor item;

identifying a routine of a plurality of predetermined routines that corresponds to the categorical data of the anchor item, each routine of the plurality of predetermined routines being (i) based on an item purchasing behavior of the corresponding customer of the plurality of customers associated with the website and (ii) associated with a set of categorical data indicating a group of categories that are associated with the corresponding routine;

generating relevant categorical data within the identified routine by applying a ranking algorithm to the categorical data of the identified routine and the categorical data of the anchor item, the relevant categorical data indicating one or more categories that are associated with the identified routine and are relevant to the anchor item, the one or more categories being associated with one or more relevant items;

generating relevant item data from the relevant categorical data by applying an item recommendation model to item data that corresponds to the relevant categorical data; and

based on the relevant item data, generating, by the processor and for each of the one or more relevant items, a graphical representation and a selectable feature, the selectable feature, when triggered, implements a set of operations that add the corresponding relevant item of the one or more relevant items to an online shopping cart associated with the website and the corresponding customer, along with the anchor item.

15. The computer program product of claim 14 , wherein the program instructions further comprise:

applying a word embedding model to purchase history data and descriptive information data of items purchased by a customer to generate sets of embeddings;

grouping the sets of embeddings together to generate categories of items; and

for each cluster of a group of categories that have a complementary behavior, determine a routine.

16. The computer program product of claim 14 , wherein the program instructions further comprise:

applying the item recommendation model to output complementary scores for items corresponding to the item data; and

ranking the items within the category based on relevancy to the anchor item that corresponds to the anchor item data.