liveBook · Manning

Data modality

As explained in Fight Fraud with Machine Learning

Data modality is the form in which data is represented, such as structured, unstructured, or semi-structured data.

Data modality is the form in which data is represented. In fraud detection, identifying the modality helps determine what evidence is available, which models fit the problem, and where deployment constraints matter. section 1.5

Common modalities used to describe fraud-detection data include:

  • Structured data: tabular values such as decimal and binary fields. section 1.5
  • Unstructured data: natural text, images, or image-only document evidence. section 1.5 section 1.5.4
  • Semi-structured data: evidence combining elements of structured and unstructured data. section 1.5

Digital document forgery

Forged passports, driving licenses, or bills rely on image manipulation. Because the evidence consists only of images, detection uses computer vision and deep learning to identify distortions across pixels. section 1.5.4

Figure 1.14 Three ways of categorizing online fraud: the industry or sector in which fraud is perpetrated, the nature of the fraudulent attack, and the modalities of data involved.

Use the categorization figure to compare data modality with the industry and attack-nature lenses for analyzing online fraud. section 1.5

How Fight Fraud with Machine Learning covers it

How deep it goes
Classifies structured, unstructured, and semi-structured data, then applies the distinction to image-based document forgery.
Examples
Uses transaction fields, text and images, combined phishing evidence, and forged digital medical bills as examples.
What you can do after
Evaluate fraud data by representation type and select detection approaches aligned with its modality.
How it connects
Connects fraud detection with structured data, unstructured data, semi-structured data, computer vision, and deep learning.

Also explained in

Compare how 5 books explain data modality →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Fight Fraud with Machine Learning. The section links point to the parts of the book it draws on.