Made some improvements to the smallsim_hard example.

analysis/smallsim_hard.Rmd

@@ -1,5 +1,5 @@
 ---
-title: Toy example comparing EM vs. SCD (hard case)
+title: Toy example comparing EM vs. SCD
 author: Peter Carbonetto
 output: workflowr::wflow_html
 ---
@@ -40,15 +40,13 @@ simulate_correlated_loadings <- function (n, k) {
   }
   return(normalize.rows(L))
 }
-set.seed(1)
+set.seed(2)
 n <- 100
 m <- 400
 k <- 6
 F <- simulate_factors(m,k)
-L <- simulate_loadings(n,k,S)
-s <- simulate_sizes(n)
-set.seed(1)
 L <- simulate_correlated_loadings(n,k)
+s <- simulate_sizes(n)
 X <- simulate_multinom_counts(L,F,s)
 X <- X[,colSums(X > 0) > 0]
 ```
@@ -67,9 +65,38 @@ ADD TEXT HERE.
 control <- list(extrapolate = FALSE,numiter = 4)
 fit0 <- fit_poisson_nmf(X,k,numiter = 20,method = "em",control = control)
 fit1 <- fit_poisson_nmf(X,fit0=fit0,numiter=180,method="em",control=control)
-fit2 <- fit_poisson_nmf(X,fit0=fit0,numiter=180,method="scd",control=control)
+fit2 <- fit_poisson_nmf(X,fit0=fit0,numiter=80,method="scd",control=control)
+control$extrapolate <- TRUE
+fit2 <- fit_poisson_nmf(X,fit0=fit2,numiter=100,method="scd",control=control)
 fit0 <- poisson2multinom(fit0)
 fit1 <- poisson2multinom(fit1)
 fit2 <- poisson2multinom(fit2)
-loadings_scatterplot(fit1$L,fit2$L,topic_colors,"em","scd")
+print(loadings_scatterplot(fit1$L,fit2$L,topic_colors,"em","scd"))
+```
+
+```{r}
+pdat <- rbind(data.frame(iter   = 1:200,
+                         ll     = fit1$progress$loglik.multinom,
+                         res    = fit1$progress$res,
+                         method = "em"),
+			  data.frame(iter   = 1:200,
+			             ll     = fit2$progress$loglik.multinom,
+						 res    = fit2$progress$res,
+						 method = "scd"))
+pdat <- transform(pdat,
+                  ll   = max(ll) - ll + 0.1)
+p <- ggplot(pdat,aes(x = iter,y = ll,color = method)) +
+  geom_line(size = 0.75) +
+  scale_y_continuous(trans = "log10") +
+  scale_color_manual(values = c("dodgerblue","darkorange")) +
+  labs(x = "iteration",y = "loglik difference") +
+  theme_cowplot(font_size = 10)
+print(p)
+```
+
+```{r}
+p1 <- simdata_structure_plot(L,topic_colors)
+p2 <- simdata_structure_plot(fit1$L,topic_colors)
+p3 <- simdata_structure_plot(fit2$L,topic_colors)
+plot_grid(p1,p2,p3,nrow = 3,ncol = 1)
 ```