Temporal Prediction

Temporal prediction involves forecasting future states or events based on patterns in time-series data. In the context of information diffusion, temporal prediction means using time-based features (timing of shares, intervals between shares, acceleration/deceleration of spread) to predict cascade growth.

Key insight from cascade research

Temporal features are often the most predictive signals for cascade growth:

• Time elapsed since original post
• Time between successive shares
• Acceleration or deceleration in sharing rate
• Views per unit time
• Number of new users per share

Related topics

• Cascade Prediction
• Information diffusion in social networks