User:Francis Tyers/Perceptron - Revision history

Francis Tyers: /* Training algorithm */

2014-11-09T10:51:20Z

Training algorithm

2014-11-09T10:42:25Z

2014-11-09T10:42:13Z

2014-11-09T10:40:03Z

2014-11-09T10:36:21Z

2014-11-09T10:35:48Z

2014-11-09T10:30:47Z

2014-11-08T22:33:37Z

2014-11-08T22:33:07Z

2014-11-08T22:26:13Z

Training data

@@ Line 18: / Line 18: @@
 The objective of the training algorithm is to find the most adequate set of weights, <math>w</math> and a threshold <math>\theta</math>.
 <div style="padding: 1em;border: 1px dashed #2f6fab;color: black;background-color: #f9f9f9;line-height: 1.1em">
 <source lang="python">
+def decision(input, weights, threshold): #{
+#}
 # Initialise weights and threshold.
 weights = [0.0, 0.0];
 threshold = 0.0;
+errors = 0;
+while True: #{
+        # If there are no errors, training has converged.
+        if errors == 0: #{
+                break;
+        #}
+#}
 </source>
 </div>

← Older revision		Revision as of 10:42, 9 November 2014
Line 20:		Line 20:


	<div style="padding: 1em;border: 1px dashed #2f6fab;color: black;background-color: #f9f9f9;line-height: 1.1em~~; font-size: 85%~~">		<div style="padding: 1em;border: 1px dashed #2f6fab;color: black;background-color: #f9f9f9;line-height: 1.1em">
	<source lang="python">		<source lang="python">

@@ Line 18: / Line 18: @@
 The objective of the training algorithm is to find the most adequate set of weights, <math>w</math> and a threshold <math>\theta</math>.
+<div style="padding: 1em;border: 1px dashed #2f6fab;color: black;background-color: #f9f9f9;line-height: 1.1em; font-size: 85%">
+<source lang="python">
+# Initialise weights and threshold.
+weights = [0.0, 0.0];
+threshold = 0.0;
+</source>
+</div>
 ==Example==

@@ Line 15: / Line 15: @@
 </math>
-==Learning algorithm==
+==Training algorithm==
+The objective of the training algorithm is to find the most adequate set of weights, <math>w</math> and a threshold <math>\theta</math>.
 ==Example==

@@ Line 12: / Line 12: @@
 <math>
-h_s(t,c) = \begin{cases}1 & \text{if } t~ =~ estacio' ~ \mathrm{and}~ sec~ \mathrm{follows}~ estacio'~ \\0 & \text{otherwise}\end{cases}
+h_{\mathrm{estacio'}}(t,c) = \begin{cases}1 & \text{if } t~ =~ season ~ \mathrm{and}~ sec~ \mathrm{follows}~ estacio'~ \\0 & \text{otherwise}\end{cases}
 </math>

@@ Line 9: / Line 9: @@
 * Threshold, <math>\theta</math>
-The input vector is made up of binary features, such as
+The input vector is made up of binary features, such as:
+<math>
+h_s(t,c) = \begin{cases}1 & \text{if } t~ =~ estacio' ~ \mathrm{and}~ sec~ \mathrm{follows}~ estacio'~ \\0 & \text{otherwise}\end{cases}
+</math>
 ==Learning algorithm==
@@ Line 17: / Line 21: @@
 ==Example==
-Here is a worked example of a perceptron applied to the task of lexical selection. Lexical selection is the task of choosing a target translation <math>t*</math> for a given source word <math>s</math> in a context <math>C</math> out of a set of possible translations <math>T</math>. A perceptron makes a classification decision for a single class, so we need to train a separate perceptron for each possible target word selection.
+Here is a worked example of a perceptron applied to the task of lexical selection. Lexical selection is the task of choosing a target translation <math>t*</math> for a given source word <math>s</math> in a context <math>c</math> out of a set of possible translations <math>T</math>. A perceptron makes a classification decision for a single class, so we need to train a separate perceptron for each possible target word selection.
 In the example,

@@ Line 5: / Line 5: @@
 The classifier consists of:
 * An input vector, <math>x</math>
 * A weight vector, <math>w</math>
 * Threshold, <math>\theta</math>
+The input vector is made up of binary features, such as
 ==Learning algorithm==

@@ Line 6: / Line 6: @@
 * Binary features
-* Weights
+* Weights, <math>w</math>
+* Threshold, <math>\theta</math>
 ==Learning algorithm==

@@ Line 43: / Line 43: @@
 {|class=wikitable
-! $c_i$ !!
+! <math>c_i</math> !!
 |-
 |-