At Mash’n Learn, all our features are built to reduce e-Commerce pain and by Research & Development. WooCommerce Product Categorization by Applying Machine Learning has been inspired by the great work of Sushant Shankar and Irving Lin from the Department of Computer Science at Stanford University.
Our Lab developed a set of functions analysing Categories from our Product Feeds partners
The functions we built are analysing the hundreds of categories from stores in Home Design and Electronics and turning them into a well designed and simplified category tree. This is mostly important for Home pages as we recommend e-Commerce owner to limit their main parent categories to maximum 10.
3 categories can be featured from the home page and 7 others in the Menu links. Above 10, we highly recommend to think about splitting their catalog in multiple shop or subsection (e.g. splitting Home and Garden into 2 separate stores).
The rationalization of categories helps the user to have a greater experience browsing a store. If it takes multiple pages and levels to find the right product, your shop’s conversion rate will struggle.
The research that inspired our WooCommerce Product Categorization feature
Applying Machine Learning to Product Categorization. Irving Lin, Sushant Shankar. [pdf]
Small to medium sized businesses who sell products online spend a significant part of their time, money, and effort organizing the products they sell, understanding consumer behavior to better market their products, and determining which products to sell. We would like to use machine learning techniques to define product categories (e.g. ‘Electronics’) and potential subcategories (e.g., ‘Printers’).
This is useful for the case where a business has a list of new products that they want to sell and they want to automatically classify these products based on training data of the businesses’ other products and classifications. This will also be useful when there is a new product line that has not been previously introduced in the market before, or the products are more densely populated than the training data (for example, if a business just sells electronic equipment, we would want to come up with a more granular structure). For this algorithm to be used in industry, we have consulted with a few small-to-medium sized companies and find that we will need an accuracy range of 95% when we have enough prior training data and a dense set of categorizations.
At Mash’n Learn, we fix e-Commerce with Machine Learning
From Natural Language Generation to Predictive Analysis, Mash’n Learn provides a complete tool suite for the large catalog retailers.
You can fill this form to have us to contact you