computeattention: Compute global attention weights and context vectors for time...

In Techtonique/ahead: Time Series Forecasting with uncertainty quantification

Compute global attention weights and context vectors for time series

Description

Compute global attention weights and context vectors for time series

Usage

computeattention(series)

Arguments

series

Numeric vector containing the time series of length n

Value

List containing:

attention_weights	n × n matrix where entry (i,j) represents the attention weight of time j on time i. Only entries j <= i are non-zero (causal attention).
context_vectors	Vector of length n where each entry i is the weighted sum of all values up to time i, using the attention weights.

Examples

# For a series of length 5
series <- c(1, 2, 3, 4, 5)
result <- computeattention(series)

# attention_weights will be 5x5 matrix
# context_vectors will be length 5
dim(result$attention_weights)  # [1] 5 5
length(result$context_vectors) # [1] 5

Techtonique/ahead documentation built on April 14, 2025, 12:51 p.m.