http://git-wip-us.apache.org/repos/asf/predictionio-site/blob/8ce20376/evaluation/paramtuning/index.html
----------------------------------------------------------------------
diff --git a/evaluation/paramtuning/index.html
b/evaluation/paramtuning/index.html
index acc21b1..229b890 100644
--- a/evaluation/paramtuning/index.html
+++ b/evaluation/paramtuning/index.html
@@ -11,11 +11,9 @@
<span class="o">}</span>
</pre></td></tr></tbody></table> </div> <h3 id='build-and-run-the-evaluation'
class='header-anchors'>Build and run the evaluation</h3><p>To run an
evaluation, the command <code>pio eval</code> is used. It takes two mandatory
parameter, 1. the <code>Evaluation</code> object, which tells PredictionIO the
engine and metric we use for the evaluation; and 2. the
<code>EngineParamsGenerator</code>, which contains a list of engine params to
test against. The following command kickstarts the evaluation workflow for the
classification template.</p><div class="highlight shell"><table
style="border-spacing: 0"><tbody><tr><td class="gutter gl" style="text-align:
right"><pre class="lineno">1
2
-3
-4</pre></td><td class="code"><pre><span class="gp">$ </span>pio build
+3</pre></td><td class="code"><pre><span class="gp">$ </span>pio build
...
-<span class="gp">$ </span>pio <span class="nb">eval
</span>org.template.classification.AccuracyEvaluation <span class="se">\</span>
- org.template.classification.EngineParamsList
+<span class="gp">$ </span>pio <span class="nb">eval
</span>org.example.classification.AccuracyEvaluation
org.example.classification.EngineParamsList
</pre></td></tr></tbody></table> </div> <p>You will see the following
output:</p><div class="highlight shell"><table style="border-spacing:
0"><tbody><tr><td class="gutter gl" style="text-align: right"><pre
class="lineno">1
2
3
@@ -102,7 +100,7 @@ Optimal Engine Params:
<span class="o">}</span>
<span class="o">}</span>
Metrics:
- org.template.classification.Accuracy: 0.9281045751633987
+ org.example.classification.Accuracy: 0.9281045751633987
The best variant params can be found <span class="k">in </span>best.json
<span class="o">[</span>INFO] <span class="o">[</span>CoreWorkflow<span
class="nv">$]</span> runEvaluation completed
</pre></td></tr></tbody></table> </div> <p>The console prints out the
evaluation metric score of each engine params, and finally pretty print the
optimal engine params. Amongst the 3 engine params we evaluate, <em>lambda =
10.0</em> yields the highest accuracy score of ~0.9281.</p><h3
id='deploy-the-best-engine-parameter' class='header-anchors'>Deploy the best
engine parameter</h3><p>The evaluation module also writes out the best engine
parameter to disk at <code>best.json</code>. We can train and deploy this
specify engine variant using the extra parameter <code>-v</code>. For
example:</p><div class="highlight shell"><table style="border-spacing:
0"><tbody><tr><td class="gutter gl" style="text-align: right"><pre
class="lineno">1
@@ -291,11 +289,9 @@ The best variant params can be found <span class="k">in
</span>best.json
<span class="o">}</span>
</pre></td></tr></tbody></table> </div> <p>A good practise is to first define
a base engine params, it contains the common parameters used in all evaluations
(lines 7 to 8). With the base params, we construct the list of engine params we
want to evaluation by adding or replacing the controller parameter. Lines 13 to
16 generate 3 engine parameters, each has a different smoothing
parameters.</p><h2 id='running-the-evaluation' class='header-anchors'>Running
the Evaluation</h2><p>It remains to run the evaluation. Let's recap the
quick start section above. The <code>pio eval</code> command kick starts the
evaluation, and the result can be seen from the console.</p><div
class="highlight shell"><table style="border-spacing: 0"><tbody><tr><td
class="gutter gl" style="text-align: right"><pre class="lineno">1
2
-3
-4</pre></td><td class="code"><pre><span class="gp">$ </span>pio build
+3</pre></td><td class="code"><pre><span class="gp">$ </span>pio build
...
-<span class="gp">$ </span>pio <span class="nb">eval
</span>org.template.classification.AccuracyEvaluation <span class="se">\</span>
- org.template.classification.EngineParamsList
+<span class="gp">$ </span>pio <span class="nb">eval
</span>org.example.classification.AccuracyEvaluation
org.example.classification.EngineParamsList
</pre></td></tr></tbody></table> </div> <p>You will see the following
output:</p><div class="highlight shell"><table style="border-spacing:
0"><tbody><tr><td class="gutter gl" style="text-align: right"><pre
class="lineno">1
2
3
@@ -390,4 +386,4 @@ The best variant params can be found <span class="k">in
</span>best.json
e=d.getElementsByTagName(t)[0];s.async=1;s.src=u;e.parentNode.insertBefore(s,e);
})(window,document,'script','//s.swiftypecdn.com/install/v1/st.js','_st');
-_st('install','HaUfpXXV87xoB_zzCQ45');</script><script
src="/javascripts/application-1a70e440.js"></script></body></html>
\ No newline at end of file
+_st('install','HaUfpXXV87xoB_zzCQ45');</script><script
src="/javascripts/application-6d62b164.js"></script></body></html>
\ No newline at end of file
