Fix Stan file for Windows

This commit is contained in:
Ben Letham 2018-05-22 09:27:00 -07:00
parent ab09fbcd6e
commit aa37fb48ac

View file

@ -1,13 +1,13 @@
functions { functions {
matrix get_changepoint_matrix(vector t, vector t_change, int T, int S) { real[ , ] get_changepoint_matrix(real[] t, real[] t_change, int T, int S) {
// Assumes t and t_change are sorted. // Assumes t and t_change are sorted.
matrix[T, S] A; real A[T, S];
row_vector[S] a_row; real a_row[S];
int cp_idx; int cp_idx;
// Start with an empty matrix. // Start with an empty matrix.
A = rep_matrix(0, T, S); A = rep_array(0, T, S);
a_row = rep_row_vector(0, S); a_row = rep_array(0, S);
cp_idx = 1; cp_idx = 1;
// Fill in each row of A. // Fill in each row of A.
@ -23,9 +23,9 @@ functions {
// Logistic trend functions // Logistic trend functions
vector logistic_gamma(real k, real m, vector delta, vector t_change, int S) { real[] logistic_gamma(real k, real m, real[] delta, real[] t_change, int S) {
vector[S] gamma; // adjusted offsets, for piecewise continuity real gamma[S]; // adjusted offsets, for piecewise continuity
vector[S + 1] k_s; // actual rate in each segment real k_s[S + 1]; // actual rate in each segment
real m_pr; real m_pr;
// Compute the rate in each segment // Compute the rate in each segment
@ -43,45 +43,49 @@ functions {
return gamma; return gamma;
} }
vector logistic_trend( real[] logistic_trend(
real k, real k,
real m, real m,
vector delta, real[] delta,
vector t, real[] t,
vector cap, real[] cap,
matrix A, real[ , ] A,
vector t_change, real[] t_change,
int S, int S,
int T int T
) { ) {
vector[S] gamma; real gamma[S];
vector[T] Y; real Y[T];
gamma = logistic_gamma(k, m, delta, t_change, S); gamma = logistic_gamma(k, m, delta, t_change, S);
for (i in 1:T) { for (i in 1:T) {
Y[i] = cap[i] / (1 + exp(-(k + dot_product(A[i], delta)) * (t[i] - (m + dot_product(A[i], gamma))))); Y[i] = cap[i] / (1 + exp(-(k + dot_product(A[i], delta))
* (t[i] - (m + dot_product(A[i], gamma)))));
} }
return Y; return Y;
} }
// Linear trend function // Linear trend function
vector linear_trend( real[] linear_trend(
real k, real k,
real m, real m,
vector delta, real[] delta,
vector t, real[] t,
matrix A, real[ , ] A,
vector t_change, real[] t_change,
int S, int S,
int T int T
) { ) {
vector[S] gamma; real gamma[S];
vector[T] Y; real Y[T];
gamma = (-t_change .* delta); for (i in 1:S) {
gamma[i] = -t_change[i] * delta[i];
}
for (i in 1:T) { for (i in 1:T) {
Y[i] = (k + dot_product(A[i], delta)) * t[i] + (m + dot_product(A[i], gamma)); Y[i] = (k + dot_product(A[i], delta)) * t[i] + (
m + dot_product(A[i], gamma));
} }
return Y; return Y;
} }
@ -90,37 +94,37 @@ functions {
data { data {
int T; // Number of time periods int T; // Number of time periods
int<lower=1> K; // Number of regressors int<lower=1> K; // Number of regressors
vector[T] t; // Time real t[T]; // Time
vector[T] cap; // Capacities for logistic trend real cap[T]; // Capacities for logistic trend
vector[T] y; // Time series real y[T]; // Time series
int S; // Number of changepoints int S; // Number of changepoints
vector[S] t_change; // Times of trend changepoints real t_change[S]; // Times of trend changepoints
matrix[T,K] X; // Regressors real X[T,K]; // Regressors
vector[K] sigmas; // Scale on seasonality prior vector[K] sigmas; // Scale on seasonality prior
real<lower=0> tau; // Scale on changepoints prior real<lower=0> tau; // Scale on changepoints prior
int trend_indicator; // 0 for linear, 1 for logistic int trend_indicator; // 0 for linear, 1 for logistic
vector[K] s_a; // Indicator of additive features real s_a[K]; // Indicator of additive features
vector[K] s_m; // Indicator of multiplicative features real s_m[K]; // Indicator of multiplicative features
} }
transformed data { transformed data {
matrix[T, S] A; real A[T, S];
A = get_changepoint_matrix(t, t_change, T, S); A = get_changepoint_matrix(t, t_change, T, S);
} }
parameters { parameters {
real k; // Base trend growth rate real k; // Base trend growth rate
real m; // Trend offset real m; // Trend offset
vector[S] delta; // Trend rate adjustments real delta[S]; // Trend rate adjustments
real<lower=0> sigma_obs; // Observation noise real<lower=0> sigma_obs; // Observation noise
vector[K] beta; // Regressor coefficients real beta[K]; // Regressor coefficients
} }
transformed parameters { transformed parameters {
vector[T] trend; real trend[T];
vector[T] Y; real Y[T];
vector[T] Xb_a; real beta_m[K];
vector[T] Xb_m; real beta_a[K];
if (trend_indicator == 0) { if (trend_indicator == 0) {
trend = linear_trend(k, m, delta, t, A, t_change, S, T); trend = linear_trend(k, m, delta, t, A, t_change, S, T);
@ -128,10 +132,15 @@ transformed parameters {
trend = logistic_trend(k, m, delta, t, cap, A, t_change, S, T); trend = logistic_trend(k, m, delta, t, cap, A, t_change, S, T);
} }
for (i in 1:K) {
beta_m[i] = beta[i] * s_m[i];
beta_a[i] = beta[i] * s_a[i];
}
for (i in 1:T) { for (i in 1:T) {
Xb_a[i] = dot_product(X[i], beta .* s_a); Y[i] = (
Xb_m[i] = dot_product(X[i], beta .* s_m); trend[i] * (1 + dot_product(X[i], beta_m)) + dot_product(X[i], beta_a)
Y[i] = trend[i] * (1 + Xb_m[i]) + Xb_a[i]; );
} }
} }