Preparing your Data for PrismRCL

PrismRCL is a sophisticated Windows-based AI classification tool designed to work with various types of data including images, text, and tabular data. To ensure optimal performance and accurate results, it’s crucial that your data is correctly organized and formatted before training begins. This guide provides essential instructions for preparing your dataset for use with PrismRCL.

Data Types Supported #

PrismRCL supports three primary types of data:

• Images: PNG format only.
• Text Data: Individual samples must be stored in separate text files.
• Tabular Data: Features should be space-separated and presented in a single line per sample in a text file. Starting from version 2.4.0, normalization of tabular data values is not required.

Required Dataset Structure #

To facilitate the AI classification process, your dataset must be organized into a specific structure. The structure is designed to segregate samples by class, making it easier for PrismRCL to understand and process the data. Here’s how to organize your dataset:

• Parent Folder: The root directory where your dataset will reside.
  • Class1 Folder: A subfolder dedicated to a specific class, containing all relevant samples for that class.
    • Images: img1.png, img2.png, etc.
    • Text Data: file1.txt, file2.txt, etc.
  • Class2 Folder, Class3 Folder, etc.: Additional subfolders for each class within your dataset, organized in the same manner as Class1 Folder.

Note: Ensure that file names across all class folders are unique to prevent any data conflicts or overwrites.

Example Dataset Structure #

This structure should be replicated for each class within your dataset, with respective samples stored in the appropriate class folders.

Preparing Your Data #

1. Sort and Classify: Begin by sorting your data samples into their respective classes. Each class should represent a category or label that you wish to train PrismRCL to recognize.
2. Organize Files: For each class, create a dedicated folder within your parent dataset folder. Name each folder appropriately to reflect the class it represents.
3. Format Samples: Ensure that all images are in PNG format, text data is in individual text files, and tabular data is formatted with space-separated values in single-line text files.
4. Check File Names: Verify that each file across your dataset has a unique name to avoid any processing errors.

Conclusion #

Proper organization and formatting of your data are key to the successful application of PrismRCL for AI classification tasks. By following the guidelines provided in this article, you can prepare your dataset efficiently and effectively, paving the way for accurate and reliable AI classification.