Amazon is the largest e-commerce and cloud computing website in the world and hosts million of products. This is a classification model that classifies 57,000 products on Amazon, using image of the product, ASIN, product title and product brand. An ensemble of classification models are used to predcit the final classes of the products.
Since the project is limited to only 63 catagories of classification, only electronic products hosted on the website are used. A combination of Tf-Learn and tensorflow are used as packages for classification. The combined classification accuracy is calculated to be 83%, possibly because the number of training samples per class is around 100, whcih is pretty less for achieving high accuracy in DNN.
- Cables
- Security & Surveillance
- Streaming Media
- Television Accessories
- Monitor Risers
- Gaming Accessories
- Video Games
- Video Cameras
- 3D Printers & Supplies
- Drones
- Mice
- Computer Accessories
- Keyboards
- Monitor Mounts
- Monitors
- Office Electronics
- Camera Accessories
- Range Extenders
- Ink & Toner
- Car & Vehicle Electronics
- Video Projectors
- Tablet Accessories
- Car Subwoofers & Amplifiers
- Tablets
- Laptop Accessories
- Tripods & Monopods
- Televisions
- Batteries
- Desktops
- Laptops
- Home Audio
- GPS & Navigation
- Radar Detectors
- Mobile Phone Accessories
- Headsets
- Binoculars & Scopes
- Modems
- Cases & Screen Protectors
- TV Mounts & Stands
- eBook Readers & Accessories
- Computer Data Storage
- Portable Audio & Speakers
- Power Management
- Computer Components
- Video Cards
- Printers & Scanners
- Memory Cards & Flash Drives
- Unlocked Cell Phones
- Wearable Technology
- Motherboards
- Telescopes
- Routers & Networking
- Car Dash Cams
- Microphones and Accessories
- Two Way Radios
- Blu-ray and DVD Players
- Standing Desks
- Cameras
- Switches
- Calculators
- Camera Lenses
- Game Consoles