Term Library

Index:

1. Introduction
2. How to access

3. How it works?

4. What are the limitations?

 

Introduction

Create rules to classify sentiment based on specific terms and optimize your research.

 

The Term Library is a solution designed for you to classify sentiment considering only a specific term within the Automation area. This functionality also allows you to create a reserve of specific terms that can be classified as positive, negative, or neutral according to your research needs.

 

By default, sentiment annotation rules made in the Automation area or in the Search Configuration are applied to the entire text of a publication. With the Term Library, you have the alternative for situations where it is necessary to apply the rule to a specific term.

 

For example: you can annotate any publication in your research with the term “chocolate” with a positive sentiment. To do this, simply include the term in the list of words associated with that sentiment. According to the previous rule, a publication with the phrase “I hate chocolate” would likely have a negative sentiment (due to the weight of the word “hate”). With the Term Library, you can associate "chocolate" with a positive sentiment and enhance this process.

 

It is important to note that the sentiment defined for a term will not necessarily be dominant in the publication it is inserted into. In other words, a publication containing the term “chocolate” has a good chance of being considered positive, but since this classification depends on the context and other words in the publication, we cannot assert that all publications with this term will always have a positive sentiment.

 

Another example: in the publication "How to Get Away with Murder”, it gets better with each episode, with the Term Library, you can add "How to Get Away with Murder" as neutral, and the publication would be considered positive, even with a term with a negative connotation like “murder”.

 

How to access?

On the Automation page, click the “Term Library” button located in the upper right corner, above the list of rules and next to the “Create” button. A new screen will open, displaying the sentiment classification options. 

 

From there, we have two possibilities: if you are a new user or have never used the previous classification rule (i.e., applied to the entire text), the sentiment rule box will already be displayed according to the new rule.

 

 

However, if you already have rules created previously, to activate the sentiment change only for the term, you will need to click the "Switch to the new behavior" button located below the explanatory text about the rule. 

 

After clicking this button, a dialog box will ask for confirmation of the change. If everything is correct, click "Confirm change". 

 

Important! This step must be done carefully, as once confirmed it will not be possible to return to the previous classification mode.

 

How it works?

Inside the sentiment rule box, you will have three fields for including terms: positive, neutral, and negative. Returning to the previous example: to include the term "chocolate" with a positive sentiment, enter it in the corresponding field and click the "+" button. The term will appear below in the list, indicating that it has been successfully included. 

 

 

After completing this step, confirm the inclusion of the term by clicking the "Save Change" button and then "OK".

 

If necessary, you can delete the inserted term by clicking the "x" icon next to the word. Within the list of terms, you will have some more options:

 

 

Clear all

Remove all terms that were listed in the desired sentiment.

Cut 

Works similarly to the "Ctrl + x" command, meaning: you can cut the listed terms to paste into a notes file or switch from one sentiment to another (for example, changing all terms registered as positive to negative).

Paste

In turn, this function is similar to the "Ctrl + v" command and allows you to paste terms copied from other lists into the desired sentiment field.

Attention! It is not possible to insert the same term into more than one sentiment. If you try to paste a word already listed in another sentiment, the following message will be displayed: 

 

What are the limitations?

Rule reprocessing in Automation does not apply to the Term Library. This means that if you make a change to the listed terms, these changes cannot be applied to previously collected publications.

There is no limit to the number of terms that can be inserted into each sentiment, but to be correctly included in your library, the listed terms must have up to 60 characters. 

 

Additionally, the system does not recognize words with punctuation (or terms containing only punctuation and space), special characters, and does not differentiate between uppercase and lowercase. When adding a new term, it is necessary to take these points into consideration.

For more information, visit the discussion on the subject at our community or videos on our channel. 😃