From the course: Introduction to AI-Native Vector Databases

Unlock the full course today

Join today to access over 24,800 courses taught by industry experts.

Structured versus unstructured data

Structured versus unstructured data

Is an image of a dog more similar to that of a cat or a wolf? Probably the wolf. Right. Can you give me a number of how similar? This one's a bit more difficult. To put a number to it and quantify how similar two images are is quite difficult. Let's explore this further. For numerical and structured data, for example, data that can be stored in an Excel file, this type of comparison is quite easy. A question could be, which customer is older, earns more, and by how much? We could simply subtract rows and it's easier for us to compare and measure structured data. The question then is, how do we compare similarity for unstructured data such as images, audio, video, or even text? Qualitatively comparing these data types is not that hard, but putting an exact number to it is. This is because structured data is easy to compare numerically, and we can use computers to perform mathematical operations on it. Computers understand and talk numbers, but unstructured data is hard to understand…

Contents