Categorize text

​

Introduction

​

Overview

​

Key Features

1. AI-powered Categorization: The Categorize Text Tool utilizes cutting-edge artificial intelligence to categorize text data. The advanced algorithms analyze the text and match it with the predefined categories, ensuring accurate and reliable categorization. Say goodbye to manual sorting and welcome the speed and accuracy of AI-powered categorization.
2. Customizable Taxonomy: Tailor the Tool to your specific needs with the customizable taxonomy feature. You can define your own list of categories or code frame by simply entering them into the Tool. This flexibility allows you to categorize text data according to your unique requirements, making the Tool adaptable to any industry or domain.
3. Tagging per Sample: Achieve more nuanced categorization with the tagging per sample feature. Specify the number of tags or categories you want to assign to each text sample, enabling you to capture the complexity and richness of your data. This feature is particularly valuable when dealing with diverse or multi-dimensional text data.
4. User-Friendly Interface: With its intuitive design, you can effortlessly input text samples and taxonomies, allowing you to focus on the analysis rather than the technicalities.

​

How to use the Tool

• Text to categorize: Enter the text sample that you wish to categorize into the “Text Input” field.
• List of categories: Specify the complete list of categories or taxonomy by entering them into the “Taxonomy” field. Each category should be on a separate line. Customize the taxonomy to align with your specific categorization needs.
  Good Categories

  1. Clear: AI is not an expert in your domain. Keep the categories simple and clear as if you are helping an intern with zero domain knowledge.
  2. Near-synonym: Hallucination is a known LLM issue. To prevent such cases, we have pushed the model to avoid self-interpretation.
  3. Meaningful combination: Use / (with space on both sides) to combine two relevant categories (e.g. Customer Service / Support)
  4. One category per concept: Avoid including overlapping categories
  5. How LLM sees/applies categories: Keep in mind that LLMs see the data under general English, not within any specific context. Categories are applied when there is a near synonym found. For instance, “Making property prices go down” is a near synonym for “Property value decline” or “Drop in value” but not a good map for “Inflation effect”.
• Maximum number of categories per Sample: Determine the number of tags or categories you want to assign to each text sample by specifying the “Tag per Sample” value. This allows for more granular categorization, capturing the nuances of your data.
• Example (Optional): Provide a few samples using samples similar to your existing data as well as matching items in your taxonomy

• Run the Tool: Once you have entered the text and defined the taxonomy, click the “Run Tool” button (on the App page) or use the run options on your data table (bulk/single run) to initiate the categorization process. Sit back and let the Tool do the heavy lifting.

  ​

  Tool execution at Relevance

  Tools and templates can be
  • tested on individually provided inputs:
    • Single run on the App page
    • Single run on the Build page
    • Single run on the data table
  • set to fetch the data from a dataset and apply the analysis on the whole dataset:
    • Bulk run on the data table
• View Results: The Tool will analyze the text and assign the appropriate categories based on the predefined taxonomy. The results will be displayed in a clear and organized format, enabling you to easily review and analyze the categorization.

​

Deep dive in the Tool

​

Tool components

​

User inputs

1. Long text input: An input text component suitable for long text pieces (more than one line), such as answers to a question, reviews, a text to summarize. This component is used twice in this Tool. Both input text (Text to categorize) and the list of categories (List of categories) are of long tex inputs.
2. Numeric input: An input component suitable for providing numeric values, such as scores, age, maximum or minimum required values. Use the default value (3) or enter your preferred maximum number of categories to be assigned to each input.
3. Table: A component for entering structured data as input, for instance, rows of samples, each containing fields such as name, last name and age. This component allows you to provide samples of text categorization done by you.
   • Enter the input sample under “text”
   • Enter matching categories under “topics”
   • Use categories/themes/topics from the list you provided with the exact same spelling
   • Use , when multiple categories/themes/topics apply

​

Tool steps

​

Properly formatting the provided categories/themes/topics

String to List formatting

​

Large Language Model (LLM)

1. Be short and precise with your instruction/request from the LLM
2. Stick to one term when referring to the same concept throughout the prompt
3. Note the goals and important instructions closes possible to the end of the prompt
4. Explicitly note constraints and goals
5. Specify a data scope using ", """ or similar identifiers
6. Include formatting instruction when necessary

​

Properly formatting the output

1. This is to make sure we can properly save the results in the desired structure.
2. Filtering out unwanted categories/themes/topics Occasionally the LLM might output categories/themes/topics that are not in the pre-specified list. This is taken care of using a simple Python code snippet in this Tool.