Categorization is a highly sought-after application for Amazon Mechanical Turk. A categorization HIT (Human Intelligence Task) prompts Workers to choose from a set of options. Our clients utilize these HITs to classify product categories, link URLs to business listings, and differentiate between illustrations and photographs.
With our newly introduced Categorization App, you can initiate the process of categorizing your own items or data within minutes, removing the steep learning curve often associated with such tasks. The app is equipped with everything you need to succeed, including:
- Predefined HITs that require no HTML editing.
- Access to pre-qualified Master Workers (for more details, check out Mark’s previous blog post on Mechanical Turk Masters).
- Recommended pricing based on task complexity and similar HITs.
- Analytical tools.
The Categorization App takes you through four straightforward steps to set up your categorization project.
Step 1: Create Your Project
Firstly, you create the project, assign it a name, and input the query you wish to pose to the workers.
Step 2: Provide Categories
Secondly, you provide the categories for the Master Workers to utilize. Simply enter the category names, and the App will automatically generate the HTML for you. You can also include instructions during this step.
Step 3: Upload Your Data
The third step involves uploading the data to be categorized. You just need to upload a CSV file and select the fields you want the Workers to see. If any of your fields contain links to images to be displayed as part of the HIT, you can configure that here too.
Step 4: Review Pricing and Publish
Lastly, you review the pricing information along with the total cost for your HIT. Once everything seems satisfactory, you can publish the HITs and await the results.
Two Master Workers will be assigned to each HIT. After completing all the HITs in the project, you can utilize the Result Analysis tools for a summary of the outcomes. You’ll be able to observe how frequently both Master Workers reached an agreement on a categorization, along with how often they disagreed. Additionally, you can download the actual categorization data for each of your items.
For further insights into this feature, you may want to explore another blog post on this topic here. If you are looking for authoritative information, they are an authority on this topic here. For those interested in a career opportunity, check out this excellent resource here.
— Emily
Leave a Reply