Sentiment Games

Four skills explore the use of sentiment analysis in Alexa skills, each provided by a different sentiment analysis service. Most sentiment analysis is used to evaluate the sentiment of an utterance after the fact. In these skills, the user is given target sentiment measurement goals, and scored based on how close the user gets to meeting those goals.

All Skills
In each of these skills, you will try to match sentiment measurements to specific goals. The closer you match the goals, the higher your score. You may not always agree with the measurements. That’s OK! And if you want to consider it a toy instead of a game, just ignore the score. That’s OK, too!

In a practice round, you can get an idea of how the service rates your statements, assigning a sentiment measurement score.

In the first round of game play, you will be asked to make a statement to match each type of sentiment.
In round one, points are awarded based on how many attempts it takes to match the sentiment.

In the second round, you will be asked to match both the predominant sentiment (e.g. POSITIVE) as well as the scored degree of that sentiment (e.g. POSITIVE @ 83%). In this round, points are awarded based on how close you get to the percentage, as well as a degree of difficulty (lower percentages are harder).

Your high score is saved, so in subsequent visits, you can try to top your previous high score.

Sentiment Match
Sentiment Match uses Amazon Comprehend for sentiment scoring, and provides separate measurements for positive, negative, neutral and mixed on a scale of 0-100 for each sentiment type.

Sentiment Blue
Sentiment Blue uses Microsoft Azure Cognitive Services for sentiment scoring, and rates your statements as positive, negative or neutral, assigning a score between 0 and 100.

Sentiment Expert
Sentiment Expert uses Expert.ai for sentiment scoring, and rates your statements as positive or negative, providing a score between 0 and 100.

Simple Sentiments
Simple Sentiments uses Symbl.ai for sentiment scoring, and rates your statements as positive, negative or neutral, assigning a single polarity score between -100 (totally negative) and 100 (totally positive) with the neutral range falling between -30 and 30.

(Icon courtesy of Vecteezy.com)