variational_fSAN is used to perform posterior inference under the finite shared atoms nested (fSAN) model with Gaussian likelihood (originally proposed in D'Angelo et al., 2023).
The model uses finite Dirichlet mixtures for both the distributional and observational levels of the model.
variational_fSAN(y, group, maxL = 30, maxK = 20,
m0 = 0, tau0 = .01, lambda0 = 3, gamma0 = 2,
alpha_bar = .005, beta_bar = .005,
epsilon = 1e-6, seed = NULL, maxSIM = 1e5,
warmstart = TRUE,verbose = FALSE)variational_fSAN returns a list of class SANvb containing four objects:
model: name of the fitted model.
params: list containing the data and the parameters used in the simulation. Details below.
sim: list containing the simulated values (optimized variational parameters). Details below.
time: total computation time.
Data and parameters:
params is a list with the following components:
y, group, Nj, JData, group labels, group frequencies, and number of groups.
K, LNumber of fitted distributional and observational clusters.
m0, tau0, lambda0, gamma0Model hyperparameters.
epsilon, seedThe threshold controlling the convergence criterion and the random seed adopted to replicate the run.
alpha_bar, beta_barthe hyperparameters governing all the finite Dirichlet distributions at the distributional and observational level.
Simulated values:
sim is a list with the following components:
theta_lMatrix of size (L,4). Each row is a posterior variational estimate of the four normal-inverse gamma hyperparameters.
Elbo_valVector containing the values of the ELBO.
XIA list of length J. Each element is a matrix of size (N, L) posterior variational probability of assignment of assignment of the i-th observation in the j-th group to the l-th OC, i.e., \(\hat{\xi}_{i,j,l} = \hat{\mathbb{Q}}(M_{i,j}=l)\).
RHOMatrix of size (J, K). Each row is a posterior variational probability of assignment of the j-th group to the k-th DC, i.e., \(\hat{\rho}_{j,k} = \hat{\mathbb{Q}}(S_j=k)\).
a_tilde_k,b_tilde_kVector of updated variational parameters of the Beta distributions governing the distributional stick-breaking process.
alpha_bar_kVector of updated variational parameters of the Dirichlet distributions governing the distributional clustering.
beta_bar_lkMatrix of updated variational parameters of the Dirichlet distributions governing the observational clustering (arranged by column).
Numerical vector of observations (required).
Numerical vector of the same length of y, indicating the group membership (required).
integers, the upper bounds for the observational and distributional clusters to fit, respectively
Hyperparameters on \((\mu, \sigma^2) \sim NIG(m_0, \tau_0, \lambda_0,\gamma_0)\).
the hyperparameter of the symmetric distributional Dirichlet distribution.
the hyperparameter of the symmetric observational Dirichlet distribution.
the tolerance that drives the convergence criterion adopted as stopping rule
random seed to control the initialization.
the maximum number of CAVI iteration to perform.
logical, if TRUE, the observational means of the cluster atoms are initialized with a k-means algorithm.
logical, if TRUE the iterations are printed.
Data structure
The finite common atoms mixture model is used to perform inference in nested settings, where the data are organized into \(J\) groups.
The data should be continuous observations \((Y_1,\dots,Y_J)\), where each \(Y_j = (y_{1,j},\dots,y_{n_j,j})\)
contains the \(n_j\) observations from group \(j\), for \(j=1,\dots,J\).
The function takes as input the data as a numeric vector y in this concatenated form. Hence y should be a vector of length
\(n_1+\dots+n_J\). The group parameter is a numeric vector of the same size as y indicating the group membership for each
individual observation.
Notice that with this specification the observations in the same group need not be contiguous as long as the correspondence between the variables
y and group is maintained.
Model
The data are modeled using a Gaussian likelihood, where both the mean and the variance are observational-cluster-specific, i.e., $$y_{i,j}\mid M_{i,j} = l \sim N(\mu_l,\sigma^2_l)$$ where \(M_{i,j} \in \{1,\dots,L \}\) is the observational cluster indicator of observation \(i\) in group \(j\). The prior on the model parameters is a Normal-Inverse-Gamma distribution \((\mu_l,\sigma^2_l)\sim NIG (m_0,\tau_0,\lambda_0,\gamma_0)\), i.e., \(\mu_l\mid\sigma^2_l \sim N(m_0, \sigma^2_l / \tau_0)\), \(1/\sigma^2_l \sim Gamma(\lambda_0, \gamma_0)\) (shape, rate).
Clustering
The model performs a clustering of both observations and groups. The clustering of groups (distributional clustering) is provided by the allocation variables \(S_j \in \{1,\dots,K\}\), with $$Pr(S_j = k \mid \dots ) = \pi_k \qquad \text{for } \: k = 1,\dots,K.$$ The distribution of the probabilities is \((\pi_1,\dots,\pi_{K})\sim Dirichlet_K(\alpha/K,\dots,\alpha/K)\). Here, the dimension \(K\) is fixed.
The clustering of observations (observational clustering) is provided by the allocation variables \(M_{i,j} \in \{1,\dots,L\}\), with $$ Pr(M_{i,j} = l \mid S_j = k, \dots ) = \omega_{l,k} \qquad \text{for } \: k = 1,\dots,K \, ; \: l = 1,\dots,L. $$ The distribution of the probabilities is \((\omega_{1,k},\dots,\omega_{L,k})\sim Dirichlet_L(\beta/L,\dots,\beta/L)\) for all \(k = 1,\dots,K\). Here, the dimension \(L\) is fixed.
D’Angelo, L., Canale, A., Yu, Z., and Guindani, M. (2023). Bayesian nonparametric analysis for the detection of spikes in noisy calcium imaging data. Biometrics, 79(2), 1370–1382. DOI: 10.1111/biom.13626
# \donttest{
set.seed(123)
y <- c(rnorm(50),rnorm(50,5))
g <- rep(1:2,rep(50,2))
est <- variational_fSAN(y, g, verbose = FALSE,
epsilon = 1e-2, maxL=15, maxK=10)
# }
Run the code above in your browser using DataLab