build the pdf

atsa-es · Jul 29, 2021 · 5e8213d · 5e8213d
1 parent 9b4d807
commit 5e8213d
Show file tree

Hide file tree

Showing 64 changed files with 844 additions and 12,002 deletions.
diff --git a/Applied_Time_Series_Analysis.Rmd b/Applied_Time_Series_Analysis.Rmd
diff --git a/Lab-fitting-multi-ss-models/multivariate-ss-ECN-NO2.Rmd b/Lab-fitting-multi-ss-models/multivariate-ss-ECN-NO2.Rmd
@@ -323,7 +323,7 @@ mod.list.corr = list(B=B, Q=Q, R=R, U=U, x0=x0, A=A, tinitx=0)
 
 Now we can fit a MARSS model and get estimates of the missing SWEs.  Convergence is slow.  We set $\mathbf{a}$ equal to the mean of the time series to speed convergence.
 
-```{r mssmiss-snotelfit-corr, results="hide"}
+```{r mssmiss-snotelfit-corr, results="hide", cache = TRUE}
 m <- apply(dat.feb, 1, mean, na.rm=TRUE)
 fit.corr <- MARSS(dat.feb, model=mod.list.corr, control=list(maxit=5000), 
                   inits=list(A=matrix(m,ns,1)))

diff --git a/Lab-fitting-multi-ss-models/multivariate-ss-missing-cov.Rmd b/Lab-fitting-multi-ss-models/multivariate-ss-missing-cov.Rmd
@@ -321,7 +321,7 @@ mod.list.corr = list(B=B, Q=Q, R=R, U=U, x0=x0, A=A, tinitx=0)
 
 Now we can fit a MARSS model and get estimates of the missing SWEs.  Convergence is slow.  We set $\mathbf{a}$ equal to the mean of the time series to speed convergence.
 
-```{r mssmiss-snotelfit-corr, results="hide"}
+```{r mssmiss-snotelfit-corr-missing, results="hide", cache = TRUE}
 m <- apply(dat.feb, 1, mean, na.rm=TRUE)
 fit.corr <- MARSS(dat.feb, model=mod.list.corr, control=list(maxit=5000), 
                   inits=list(A=matrix(m,ns,1)))

diff --git a/Lab-intro-to-jags/intro-to-jags.Rmd b/Lab-intro-to-jags/intro-to-jags.Rmd
@@ -445,8 +445,6 @@ y_t = x_t + v_t, \, v_t \sim \N(0, r)
 Because $x$ is a random walk model not a stationary AR(1), we will place a vague weakly informative prior on $x_0$: $x_0 \sim \N(y_1, 1000)$. We had to pass in `Y1` as data because JAGS would complain if we used `Y[1]` in our prior (because have `X0` in our model for $Y[1]$). `EY` is added so that we can track the model fits for $y$. In this case it is just `X` but in more complex models it will involve more parameters.
 
 ```{r jags-ss1, echo=TRUE, results='hide', cache=TRUE}
-# 5. MAKE THE SS MODEL for a stochastic level model
-
 model.loc <- ("ss_model.txt")
 jagsscript <- cat("
 model {  
@@ -836,7 +834,7 @@ removefiles <- c(removefiles, model.loc)
 There are a number of different approaches to using Bayesian time series models to perform forecasting. One approach might be to fit a model, and use those posterior distributions to forecast as a secondary step (say within R). A more streamlined approach is to do this within the JAGS code itself. We can take advantage of the fact that JAGS allows you to include NAs in the response variable (but never in the predictors). Let's use the same Wind dataset, and the univariate state-space model described above to forecast three time steps into the future. We can do this by including 3 more NAs in the dataset, and incrementing the variable ```N``` by 3.
 
 ```{r jags-cov-forecast, results='hide', cache=TRUE}
-jags.data <- list("Y" = c(Wind, NA, NA, NA), "N" = (N + 3))
+jags.data <- list("Y" = c(Wind, NA, NA, NA), "N" = (N + 3), Y1 = Wind[1])
 jags.params <- c("q", "r", "EY", "u")
 model.loc <- ("ss_model.txt")
 mod_ss_forecast <- jags(jags.data,

diff --git a/ar1_intercept.txt b/ar1_intercept.txt
diff --git a/cleanedRmd/intro-to-jags.Rmd b/cleanedRmd/intro-to-jags.Rmd
@@ -445,8 +445,6 @@ y_t = x_t + v_t, \, v_t \sim \,\text{N}(0, r)
 Because $x$ is a random walk model not a stationary AR(1), we will place a vague weakly informative prior on $x_0$: $x_0 \sim \,\text{N}(y_1, 1000)$. We had to pass in `Y1` as data because JAGS would complain if we used `Y[1]` in our prior (because have `X0` in our model for $Y[1]$). `EY` is added so that we can track the model fits for $y$. In this case it is just `X` but in more complex models it will involve more parameters.
 
 ```{r jags-ss1, echo=TRUE, results='hide', cache=TRUE}
-# 5. MAKE THE SS MODEL for a stochastic level model
-
 model.loc <- ("ss_model.txt")
 jagsscript <- cat("
 model {  
@@ -836,7 +834,7 @@ removefiles <- c(removefiles, model.loc)
 There are a number of different approaches to using Bayesian time series models to perform forecasting. One approach might be to fit a model, and use those posterior distributions to forecast as a secondary step (say within R). A more streamlined approach is to do this within the JAGS code itself. We can take advantage of the fact that JAGS allows you to include NAs in the response variable (but never in the predictors). Let's use the same Wind dataset, and the univariate state-space model described above to forecast three time steps into the future. We can do this by including 3 more NAs in the dataset, and incrementing the variable ```N``` by 3.
 
 ```{r jags-cov-forecast, results='hide', cache=TRUE}
-jags.data <- list("Y" = c(Wind, NA, NA, NA), "N" = (N + 3))
+jags.data <- list("Y" = c(Wind, NA, NA, NA), "N" = (N + 3), Y1 = Wind[1])
 jags.params <- c("q", "r", "EY", "u")
 model.loc <- ("ss_model.txt")
 mod_ss_forecast <- jags(jags.data,

diff --git a/docs/Applied_Time_Series_Analysis.pdf b/docs/Applied_Time_Series_Analysis.pdf
diff --git a/docs/Applied_Time_Series_Analysis_files/figure-html/cyclic-sockeye-corrplot-1.png b/docs/Applied_Time_Series_Analysis_files/figure-html/cyclic-sockeye-corrplot-1.png
diff --git a/docs/Applied_Time_Series_Analysis_files/figure-html/jags-marss1-hist-1.png b/docs/Applied_Time_Series_Analysis_files/figure-html/jags-marss1-hist-1.png
diff --git a/docs/Applied_Time_Series_Analysis_files/figure-html/jags-marss1-plot-1.png b/docs/Applied_Time_Series_Analysis_files/figure-html/jags-marss1-plot-1.png
diff --git a/docs/Applied_Time_Series_Analysis_files/figure-html/stan-dfa-plot-trends-1.png b/docs/Applied_Time_Series_Analysis_files/figure-html/stan-dfa-plot-trends-1.png
diff --git a/docs/Applied_Time_Series_Analysis_files/figure-html/stan-fig-burnin-1.png b/docs/Applied_Time_Series_Analysis_files/figure-html/stan-fig-burnin-1.png
diff --git a/docs/Applied_Time_Series_Analysis_files/figure-html/stan-fig-lm-1.png b/docs/Applied_Time_Series_Analysis_files/figure-html/stan-fig-lm-1.png
diff --git a/docs/Applied_Time_Series_Analysis_files/figure-html/stan-hist-1.png b/docs/Applied_Time_Series_Analysis_files/figure-html/stan-hist-1.png
diff --git a/docs/Applied_Time_Series_Analysis_files/figure-html/stan-plot-seal-1.png b/docs/Applied_Time_Series_Analysis_files/figure-html/stan-plot-seal-1.png
diff --git a/docs/Rcode/intro-to-jags.R b/docs/Rcode/intro-to-jags.R
@@ -258,8 +258,6 @@ mod_ar1_intercept <- R2jags::jags(jags.data,
 
 
 ## ----jags-ss1, echo=TRUE, results='hide', cache=TRUE-------------------------------------------
-# 5. MAKE THE SS MODEL for a stochastic level model
-
 model.loc <- ("ss_model.txt")
 jagsscript <- cat("
 model {  
@@ -517,7 +515,7 @@ mod_ss <- jags(jags.data,
 
 
 ## ----jags-cov-forecast, results='hide', cache=TRUE---------------------------------------------
-jags.data <- list("Y" = c(Wind, NA, NA, NA), "N" = (N + 3))
+jags.data <- list("Y" = c(Wind, NA, NA, NA), "N" = (N + 3), Y1 = Wind[1])
 jags.params <- c("q", "r", "EY", "u")
 model.loc <- ("ss_model.txt")
 mod_ss_forecast <- jags(jags.data,

diff --git a/docs/Rmds/intro-to-jags.Rmd b/docs/Rmds/intro-to-jags.Rmd
@@ -445,8 +445,6 @@ y_t = x_t + v_t, \, v_t \sim \,\text{N}(0, r)
 Because $x$ is a random walk model not a stationary AR(1), we will place a vague weakly informative prior on $x_0$: $x_0 \sim \,\text{N}(y_1, 1000)$. We had to pass in `Y1` as data because JAGS would complain if we used `Y[1]` in our prior (because have `X0` in our model for $Y[1]$). `EY` is added so that we can track the model fits for $y$. In this case it is just `X` but in more complex models it will involve more parameters.
 
 ```{r jags-ss1, echo=TRUE, results='hide', cache=TRUE}
-# 5. MAKE THE SS MODEL for a stochastic level model
-
 model.loc <- ("ss_model.txt")
 jagsscript <- cat("
 model {  
@@ -836,7 +834,7 @@ removefiles <- c(removefiles, model.loc)
 There are a number of different approaches to using Bayesian time series models to perform forecasting. One approach might be to fit a model, and use those posterior distributions to forecast as a secondary step (say within R). A more streamlined approach is to do this within the JAGS code itself. We can take advantage of the fact that JAGS allows you to include NAs in the response variable (but never in the predictors). Let's use the same Wind dataset, and the univariate state-space model described above to forecast three time steps into the future. We can do this by including 3 more NAs in the dataset, and incrementing the variable ```N``` by 3.
 
 ```{r jags-cov-forecast, results='hide', cache=TRUE}
-jags.data <- list("Y" = c(Wind, NA, NA, NA), "N" = (N + 3))
+jags.data <- list("Y" = c(Wind, NA, NA, NA), "N" = (N + 3), Y1 = Wind[1])
 jags.params <- c("q", "r", "EY", "u")
 model.loc <- ("ss_model.txt")
 mod_ss_forecast <- jags(jags.data,

diff --git a/docs/analysis-goal.html b/docs/analysis-goal.html
@@ -626,7 +626,6 @@ <h2><span class="header-section-number">16.2</span> Analysis goal</h2>
 <li>The amplitude of the cycle changes over time.</li>
 </ol>
 <p>Both of these will cause us problems when we try to estimate a stochastic level.</p>
-<pre><code>Warning: Removed 3 row(s) containing missing values (geom_path).</code></pre>
 <p><img src="Applied_Time_Series_Analysis_files/figure-html/unnamed-chunk-59-1.png" width="80%" style="display: block; margin: auto;" /></p>
 <p>ACF of the spawners showing the 5-year cycle in most of the rivers.
 <img src="Applied_Time_Series_Analysis_files/figure-html/unnamed-chunk-60-1.png" width="80%" style="display: block; margin: auto;" /></p>

diff --git a/docs/chap-cyclic-sockeye.html b/docs/chap-cyclic-sockeye.html
@@ -623,9 +623,9 @@ <h1><span class="header-section-number">Chapter 16</span> Modeling cyclic sockey
 <p>A script with all the R code in the chapter can be downloaded <a href="./Rcode/cyclic-sockeye.R">here</a>. The Rmd for this chapter can be downloaded <a href="./Rmds/cyclic-sockeye.Rmd">here</a></p>
 <div id="data-and-packages-10" class="section level3 unnumbered">
 <h3>Data and packages</h3>
-<div class="sourceCode" id="cb732"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb732-1"><a href="chap-cyclic-sockeye.html#cb732-1" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(atsalibrary)</span>
-<span id="cb732-2"><a href="chap-cyclic-sockeye.html#cb732-2" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(ggplot2)</span>
-<span id="cb732-3"><a href="chap-cyclic-sockeye.html#cb732-3" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(MARSS)</span></code></pre></div>
+<div class="sourceCode" id="cb720"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb720-1"><a href="chap-cyclic-sockeye.html#cb720-1" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(atsalibrary)</span>
+<span id="cb720-2"><a href="chap-cyclic-sockeye.html#cb720-2" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(ggplot2)</span>
+<span id="cb720-3"><a href="chap-cyclic-sockeye.html#cb720-3" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(MARSS)</span></code></pre></div>
 </div>
 </div>
             </section>

diff --git a/docs/chap-stan.html b/docs/chap-stan.html
@@ -625,23 +625,23 @@ <h1><span class="header-section-number">Chapter 13</span> Stan for Bayesian time
 <div id="data-and-packages-9" class="section level3 unnumbered">
 <h3>Data and packages</h3>
 <p>You will need the <strong>atsar</strong> and <strong>bayesdfa</strong> packages we have written for fitting state-space time series models with Stan. Install using the <strong>devtools</strong> package.</p>
-<div class="sourceCode" id="cb644"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb644-1"><a href="chap-stan.html#cb644-1" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(devtools)</span>
-<span id="cb644-2"><a href="chap-stan.html#cb644-2" aria-hidden="true" tabindex="-1"></a><span class="co"># Windows users will likely need to set this</span></span>
-<span id="cb644-3"><a href="chap-stan.html#cb644-3" aria-hidden="true" tabindex="-1"></a><span class="co"># Sys.setenv(&#39;R_REMOTES_NO_ERRORS_FROM_WARNINGS&#39; = &#39;true&#39;)</span></span>
-<span id="cb644-4"><a href="chap-stan.html#cb644-4" aria-hidden="true" tabindex="-1"></a>devtools<span class="sc">::</span><span class="fu">install_github</span>(<span class="st">&quot;nwfsc-timeseries/atsar&quot;</span>)</span>
-<span id="cb644-5"><a href="chap-stan.html#cb644-5" aria-hidden="true" tabindex="-1"></a>devtools<span class="sc">::</span><span class="fu">install_github</span>(<span class="st">&quot;nwfsc-timeseries/tvvarss&quot;</span>)</span>
-<span id="cb644-6"><a href="chap-stan.html#cb644-6" aria-hidden="true" tabindex="-1"></a>devtools<span class="sc">::</span><span class="fu">install_github</span>(<span class="st">&quot;fate-ewi/bayesdfa&quot;</span>)</span></code></pre></div>
+<div class="sourceCode" id="cb640"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb640-1"><a href="chap-stan.html#cb640-1" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(devtools)</span>
+<span id="cb640-2"><a href="chap-stan.html#cb640-2" aria-hidden="true" tabindex="-1"></a><span class="co"># Windows users will likely need to set this</span></span>
+<span id="cb640-3"><a href="chap-stan.html#cb640-3" aria-hidden="true" tabindex="-1"></a><span class="co"># Sys.setenv(&#39;R_REMOTES_NO_ERRORS_FROM_WARNINGS&#39; = &#39;true&#39;)</span></span>
+<span id="cb640-4"><a href="chap-stan.html#cb640-4" aria-hidden="true" tabindex="-1"></a>devtools<span class="sc">::</span><span class="fu">install_github</span>(<span class="st">&quot;nwfsc-timeseries/atsar&quot;</span>)</span>
+<span id="cb640-5"><a href="chap-stan.html#cb640-5" aria-hidden="true" tabindex="-1"></a>devtools<span class="sc">::</span><span class="fu">install_github</span>(<span class="st">&quot;nwfsc-timeseries/tvvarss&quot;</span>)</span>
+<span id="cb640-6"><a href="chap-stan.html#cb640-6" aria-hidden="true" tabindex="-1"></a>devtools<span class="sc">::</span><span class="fu">install_github</span>(<span class="st">&quot;fate-ewi/bayesdfa&quot;</span>)</span></code></pre></div>
 <p>In addition, you will need the <strong>rstan</strong>, <strong>datasets</strong>, <strong>parallel</strong> and <strong>loo</strong> packages. After installing, if needed, load the packages:</p>
-<div class="sourceCode" id="cb645"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb645-1"><a href="chap-stan.html#cb645-1" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(atsar)</span>
-<span id="cb645-2"><a href="chap-stan.html#cb645-2" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(rstan)</span>
-<span id="cb645-3"><a href="chap-stan.html#cb645-3" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(loo)</span></code></pre></div>
+<div class="sourceCode" id="cb641"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb641-1"><a href="chap-stan.html#cb641-1" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(atsar)</span>
+<span id="cb641-2"><a href="chap-stan.html#cb641-2" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(rstan)</span>
+<span id="cb641-3"><a href="chap-stan.html#cb641-3" aria-hidden="true" tabindex="-1"></a><span class="fu">library</span>(loo)</span></code></pre></div>
 <p>Once you have Stan and <strong>rstan</strong> installed, optimize Stan on your machine:</p>
-<div class="sourceCode" id="cb646"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb646-1"><a href="chap-stan.html#cb646-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rstan_options</span>(<span class="at">auto_write =</span> <span class="cn">TRUE</span>)</span>
-<span id="cb646-2"><a href="chap-stan.html#cb646-2" aria-hidden="true" tabindex="-1"></a><span class="fu">options</span>(<span class="at">mc.cores =</span> parallel<span class="sc">::</span><span class="fu">detectCores</span>())</span></code></pre></div>
+<div class="sourceCode" id="cb642"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb642-1"><a href="chap-stan.html#cb642-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rstan_options</span>(<span class="at">auto_write =</span> <span class="cn">TRUE</span>)</span>
+<span id="cb642-2"><a href="chap-stan.html#cb642-2" aria-hidden="true" tabindex="-1"></a><span class="fu">options</span>(<span class="at">mc.cores =</span> parallel<span class="sc">::</span><span class="fu">detectCores</span>())</span></code></pre></div>
 <p>For this lab, we will use a data set on air quality in New York from the <strong>datasets</strong> package. Load the data and create a couple new variables for future use.</p>
-<div class="sourceCode" id="cb647"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb647-1"><a href="chap-stan.html#cb647-1" aria-hidden="true" tabindex="-1"></a><span class="fu">data</span>(airquality, <span class="at">package =</span> <span class="st">&quot;datasets&quot;</span>)</span>
-<span id="cb647-2"><a href="chap-stan.html#cb647-2" aria-hidden="true" tabindex="-1"></a>Wind <span class="ot">&lt;-</span> airquality<span class="sc">$</span>Wind  <span class="co"># wind speed</span></span>
-<span id="cb647-3"><a href="chap-stan.html#cb647-3" aria-hidden="true" tabindex="-1"></a>Temp <span class="ot">&lt;-</span> airquality<span class="sc">$</span>Temp  <span class="co"># air temperature</span></span></code></pre></div>
+<div class="sourceCode" id="cb643"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb643-1"><a href="chap-stan.html#cb643-1" aria-hidden="true" tabindex="-1"></a><span class="fu">data</span>(airquality, <span class="at">package =</span> <span class="st">&quot;datasets&quot;</span>)</span>
+<span id="cb643-2"><a href="chap-stan.html#cb643-2" aria-hidden="true" tabindex="-1"></a>Wind <span class="ot">&lt;-</span> airquality<span class="sc">$</span>Wind  <span class="co"># wind speed</span></span>
+<span id="cb643-3"><a href="chap-stan.html#cb643-3" aria-hidden="true" tabindex="-1"></a>Temp <span class="ot">&lt;-</span> airquality<span class="sc">$</span>Temp  <span class="co"># air temperature</span></span></code></pre></div>
 </div>
 </div>
             </section>