data sets

In statistics

Numerical data sets: involve only numbers.
Bivariate data sets: Include two data variables.
Multivariate data sets: include three or more data variables.
Categorical data sets: consist of categorical variables that can only take a limited set of values.
Correlation data sets: Contain data variables that are related to each other.

Machine learning

ML Training Datasets: Used to train the model.
Data sets for validation: used to reduce overfitting and make the model more accurate.
Test data sets: These are used to test the final output of the model and confirm its accuracy.

Tips for creating image datasets.

Examples of tools:

1. Choose the appropriate annotation tool.
Labelimg – This is a free and open source image annotation tool available on three platforms: Windows, macOS, and Linux. It is written in Python and uses Qt for its graphical interface, the annotations are saved as XML files in PASCAL VOC format, the format used by ImageNet (an image database organized according to the nominal WordNet hierarchy, with hundreds and thousands of images for each Node. It has played an important role in the progress of computer vision and deep learning research. Researchers can access the data free of charge for non-commercial purposes). LabelImg also supports YOLO and CreateML formats.

VIA (VGG Image Annotator) :Self-contained, open source, easy-to-use software that allows manual annotation of images, audio, and video. It can be accessed through a web browser with no installation or configuration required. The entire program is contained in a single HTML page less than 400 kilobytes in size and can be used offline in most modern web browsers. VIA is based solely on HTML, JavaScript and CSS, and does not require external libraries. Released under the BSD-2 clause license, it is the preferred choice of many annotation services as it is suitable for both academic research and commercial applications. Available on Windows, macOS and Linux.

LabelMe – Online annotation tool provided by the MIT CSAIL team to create image databases for computer vision research. Also available for free on Windows, macOS and Linux.

Note: You can find a version on GitHub for polygon annotation.

Head: This is a commercial image annotation tool.
V7 – A commercial image annotation tool.

Advantages of using a data set

Below are the three most important advantages of using data sets

Improved decision making

The information contained in the data sets can be used to support strategic decisions. Specifically, data sets allow you to detect market trends, analyze customer behavior, identify patterns and relationships in the data, and measure performance. You can then leverage data sets to make evidence- and data-driven decisions, helping your company understand where to allocate resources, how to develop new products, and how much to charge for new services. As a result, you will improve your competitive advantage and your ability to respond to market needs.

Better user experience

Data sets containing user reviews can help you understand how to improve the overall customer experience. For example, you can use this information to create personalized experiences, improve product design, adapt or add new features, and optimize the user journey. By offering a better user experience, you will increase customer satisfaction.

Saving time and costs

You can use a data set to discover time and cost saving opportunities. For example, data sets can help identify inefficiencies in the development process, allowing you to streamline operations, reduce waste, and save time. Likewise, data sets can be explored to uncover redundant processes, business areas that spend more than necessary, and inefficiencies in the supply chain, helping to reduce costs.