-
susiepca.sim.generate_sim(seed: int, l_dim: int, n_dim: int, p_dim: int, z_dim: int, effect_size: float =
1.0) SimulatedData[source] - Create the function to generate a sparse data for PCA. Please note that to
illustrate how SuSiE PCA work, we build this simple and straitforward simulation where each latent component have exact l_dim number of non-overlapped single effects. Please make sure l_dim < p_dim/z_dim when generate simulation data using this function.
- Parameters:
- seed: int¶
Seed for “random” initialization
- l_dim: int¶
Number of single effects in each factor
- n_dim: int¶
Number of sample in the data
- p_dim: int¶
Number of feature in the data
- z_dim: int¶
Number of Latent dimensions
- effect_size: float =
1.0¶ The effect size of features contributing to the factor. (default = 1).
- Returns:
Tuple that contains simulated factors (
N x K), W (factor loadings (K x P), and data X (data (N x P).- Return type:
Last update:
Feb 13, 2024