Viewing file:  ar_sentiment.php (6.04 KB) -rw-rw-rw-
Select action/file-type:
 (+) |  (+) |  (+) | Code (+) | Session (+) |  (+) | SDB (+) |  (+) |  (+) |  (+) |  (+) |  (+) |
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title>Arabic Sentiment Analysis</title>
<meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
<link rel="stylesheet" type="text/css" href="css/style.css" media="all" />
</head>
<body>
<div class="Paragraph">
<h2>Arabic Sentiment Analysis:</h2>
<p>Sentiment analysis is one of the most useful Natural Language Processing (NLP) functionalities
that can determine the tone (positive, negative) of the text (e.g., product reviews, comments, etc.).</p>
<p>This Machine Learning (ML) model was built using a dataset published on
<a href="https://www.kaggle.com/abedkhooli/arabic-100k-reviews" target="_blank">Kaggle</a>
and combines 100k Arabic reviews from hotels, books, movies, products, and a few airlines.
Text (reviews) were cleaned by removing Arabic diacritics and non-Arabic characters.
Predictions are calculated using the log-odds statistics, and method accuracy exceeds 75%
which is not a bad performance for a model sized 28.2 KB.</p>
<i>It has been tested also with the <a href="" target="_blank">HARD: Hotel Arabic-Reviews Dataset</a>,
and it was able to achieve 82% on the balanced reviews dataset (in total 105,698 reviews).</i>
<p>For simplicity, we assumed that all the words in the first language spoken by the Semitic peoples consisted
of bi-radicals (i.e., two sounds/letters). Therefore, we can handle the majority of Arabic word roots as being
expanded by the addition of a third letter, with the resulting meaning having a semantic relation to the original bi-radical
(Ref: <a href="https://repositories.lib.utexas.edu/handle/2152/13246" target="_blank">The biradical origin of semitic roots</a>).</p>
</div><br />
<div class="Paragraph">
<h2 dir="ltr" id="example-1">
<a href="#example-1" class="anchor"><img src="./images/link_icon.png" width="16" border="0"></a>Example Output 1:</h2>
<?php
error_reporting(E_STRICT);
/*
// Autoload files using Composer autoload
require_once __DIR__ . '/../vendor/autoload.php';
*/
require '../src/Arabic.php';
$Arabic = new \ArPHP\I18N\Arabic();
$reviews = array('الخدمة كانت بطيئة',
'الإطلالة رائعة والطعام لذيذ',
'التبريد لا يعمل والواي فاي ضعيفة',
'النظافة مميزة وموظفي الاستقبال متعاونين',
'جاءت القطعة مكسورة والعلبة مفتوحة',
'المنتج مطابق للمواصفات والتسليم سريع');
echo <<< END
<center>
<table border="0" cellspacing="2" cellpadding="5" width="60%">
<tr>
<td bgcolor="#27509D" align="center" width="50%">
<b><font color="#ffffff">Arabic Review (sample input)</font></b>
</td>
<td bgcolor="#27509D" align="center" width="25%">
<b><font color="#ffffff">Sentiment (auto generated)</font></b>
</td>
<td bgcolor="#27509D" align="center" width="25%">
<b><font color="#ffffff">Probability (auto generated)</font></b>
</td>
</tr>
END;
foreach ($reviews as $review) {
$analysis = $Arabic->arSentiment($review);
if ($analysis['isPositive']) {
$sentiment = 'Positive';
$bgcolor = '#E0F0FF';
} else {
$sentiment = 'Negative';
$bgcolor = '#FFF0FF';
}
$probability = sprintf('%0.1f', round(100 * $analysis['probability'], 1));
echo '<tr><td bgcolor="'.$bgcolor.'" align="right">';
echo '<font face="Tahoma">'.$review.'</font></td>';
echo '<td bgcolor="'.$bgcolor.'" align="center">'.$sentiment.'</td>';
echo '<td bgcolor="'.$bgcolor.'" align="center">'.$probability.'%</td></tr>';
}
echo '</table></center>';
?>
</div><br />
<div class="Paragraph">
<h2 dir="ltr">Example Code 1:</h2>
<?php
$code = <<< ENDALL
<?php
\$Arabic = new \\ArPHP\\I18N\\Arabic();
\$reviews = array('الخدمة كانت بطيئة',
'الإطلالة رائعة والطعام لذيذ',
'التبريد لا يعمل والواي فاي ضعيفة',
'النظافة مميزة وموظفي الاستقبال متعاونين',
'جاءت القطعة مكسورة والعلبة مفتوحة',
'المنتج مطابق للمواصفات والتسليم سريع');
echo <<< END
<center>
<table border="0" cellspacing="2" cellpadding="5" width="60%">
<tr>
<td bgcolor="#27509D" align="center" width="50%">
<b><font color="#ffffff">Arabic Review (sample input)</font></b>
</td>
<td bgcolor="#27509D" align="center" width="25%">
<b><font color="#ffffff">Sentiment (auto generated)</font></b>
</td>
<td bgcolor="#27509D" align="center" width="25%">
<b><font color="#ffffff">Probability (auto generated)</font></b>
</td>
</tr>
END;
foreach (\$reviews as \$review) {
\$analysis = \$Arabic->arSentiment(\$review);
if (\$analysis['isPositive']) {
\$sentiment = 'Positive';
\$bgcolor = '#E0F0FF';
} else {
\$sentiment = 'Negative';
\$bgcolor = '#FFF0FF';
}
\$probability = sprintf('%0.1f', round(100 * \$analysis['probability'], 1));
echo '<tr><td bgcolor="'.\$bgcolor.'" align="right">';
echo '<font face="Tahoma">'.\$review.'</font></td>';
echo '<td bgcolor="'.\$bgcolor.'" align="center">'.\$sentiment.'</td>';
echo '<td bgcolor="'.\$bgcolor.'" align="center">'.\$probability.'%</td></tr>';
}
echo '</table></center>';
ENDALL;
highlight_string($code);
?>
<hr/><i>Related Documentation:
<a href="https://khaled-alshamaa.github.io/ar-php/classes/ArPHP-I18N-Arabic.html#method_arSentiment" target="_blank">arSentiment</a>
</i>
</div>
<footer><i><a href="https://github.com/khaled-alshamaa/ar-php">Ar-PHP</a>, an open-source library for website developers to process Arabic content</i></footer>
</body>
</html>
|