Laravel Sentiment Analysis

Introduction

A Laravel wrapper for phpInsight.

Installation

PHP 5.4+ or HHVM 3.3+, and Composer are required.

To get the latest version of Laravel Sentiment Analysis, simply add the following line to the require block of your composer.json file:

"antoineaugusti/laravel-sentiment-analysis": "~2.0"

You'll then need to run composer install or composer update to download it and have the autoloader updated.

This package supports auto discovery for Laravel 5.5+. If you don't have a version above 5.5 yet, you need to register the service provider. Open up config/app.php and add the following to the providers key.

Antoineaugusti\LaravelSentimentAnalysis\LaravelSentimentAnalysisServiceProvider::class,

You can register the SentimentAnalysis facade in the aliases key of your config/app.php file if you like.

'SentimentAnalysis' => Antoineaugusti\LaravelSentimentAnalysis\SentimentAnalysis::class,

Looking for a Laravel 4 compatible version?

Checkout the 1.2 version, installable by requiring "antoineaugusti/laravel-sentiment-analysis": "1.2".

Usage

Sentences can be classified as negative, neutral or positive. The only supported language for the moment is English.

Custom Dictionary

You can provide a custom dictionary by providing the path the folder when you create a new SentimentAnalysis object.

$analysis = new SentimentAnalysis(storage_path('custom_dictionary/'));

Please look at the PHPInsight data files to see how you should name and structure your files.

SentimentAnalysis::isNegative($sentence)

Returns a boolean telling if the given $sentence is classified as negative.

SentimentAnalysis::isNeutral($sentence)

Returns a boolean telling if the given $sentence is classified as neutral.

SentimentAnalysis::isPositive($sentence)

Returns a boolean telling if the given $sentence is classified as positive.

SentimentAnalysis::decision($sentence)

Get the sentiment of a sentence. Will return negative, neutral or positive

SentimentAnalysis::score($sentence)

Get the confidence of a decision for a result. The closer to 1, the better!

SentimentAnalysis::scores($sentence)

Get the score value for each decision. Returns an array. The closer to 1, the better! Return example:

['negative' => 0.5, 'neutral' => 0.25, 'positive' => 0.25]