✨ 👚 👕 👗 ✨
Recommending similar product based on text features.
- The use of recommendation systems is now very common in e-commerce companies like Amazon, Flipkart, Myntra, etc.
- In this project, "Bag of Words", "TF-IDF" and "Word2Vec" models are built to recommend a product based on product title text.
- Many people use e-commerce platforms to buy products. When a product is selected, people expect to see similar products.
- A model which can recommend people similar products is our Business Objective.
- Low Latency - In real time within in few nano seconds the model should be able to recommend similar product to people.
- Speed - We need speedy model not 100% accurate model. So accuracy is not important.
- Same product recommendation error - People doesn't like to see same product again and again.
- Interpretablity is important.
- Dataset contains ladies tops fashion of amazon website, initially data has 183k rows and 19 features.
- By cleaning data we brought down the number of data points from 183K to 25K.
- Of these 19 features, we will use only 7 features in this project.
- The description of this dataset is as following:
| No. | Column name | Description |
|---|---|---|
| 1 | asin | Amazon Standard Identification Number |
| 2 | brand | Brand name of the product |
| 3 | medium_image_url | URL of the product image |
| 4 | product_type_name | Type of the product |
| 5 | color | Color information of the product |
| 6 | title | Title of the product |
| 7 | formatted_price | Price of the product in ($) US Dollar |
- Titles fairly describe what the product is.
- We will use the title feature to recommend a product because they are short and informative.