TECH | Mar 6, 2018

Name of the medicine: Sentiment Analysis

If Sentiment Analysis were a medicine: read the warnings and instructions for use carefully

WHAT IT IS

According to Wikipedia, Sentiment Analysis refers to “the use of natural language processingtext analysis and computational linguistics to identify and extract subjective information from different sources”.

As the name suggests, therefore, Sentiment Analysis is “the computational analysis of feelings and opinions expressed in texts generated on the Net on a product, service, individual, organization, event or other”. It is a technique which, thanks to data mining and natural language processing, allows – or should allow – understanding of the emotions that the user has experienced in certain contexts on the basis of the reading and “automatic” interpretation of the texts of messages left online, for example on social networks.

WHY IT IS USED

The medicine is used to extract and analyze information that emerge from comments, posts, texts on social media or from other sources, usually referring to opinions expressed by users regarding a product, service or topic of discussion. Sentiment Analysis helps to identify the polarity of a certain text, classifying it as positive, negative or neutral.

In the company it is used, for example, in the context of reputation management, to understand the strengths and weaknesses of a product, service or even a promotional campaign, in order to avoid crises and to orient communication strategies thanks to the analysis of user reactions.

WHEN IT SHOULD BE USED

The medicine should be used when there is the need to detect, including in real time, the “feeling” of users on the network with regard to a certain event, happening, product or other in order to exploit the feedback to modify, transform and improve what is being analyzed

Sentiment Analysis can be used in cases of online reputation verification, in the study of the perception of a brand, product and character, and in measuring the return on web marketing activities.

There are several sectors of application, ranging from marketing to communication, politics, the study of financial markets, medicine, sport, etc.

WHAT CAN THE MEDICINE’S EFFECT CHANGE

The combination of Sentiment Analysis with active ingredients such as Big Data, Network Analysis or also Machine Learning enhances the beneficial effects. The taking of Sentiment Analysis may prove ineffective in apathetic subjects.

HOW TO USE THIS MEDICINE

How much

Sentiment Analysis should be taken on a regular basis. There are no contraindications in event of overdose.

When and for how long

Sentiment Analysis can be taken when there is information to analyze, better if in large quantities (Big Data). Sentiment Analysis can be used for a limited period in order to verify its beneficial effects in specific situations or as a long-term therapy, for constant monitoring of the context and timely identification of potential symptoms of reputational malaise.

WHAT TO DO IF YOU HAVE TAKEN AN OVERDOSE OF THE MEDICINE

Excessive doses of the medicine are harmful in subjects intolerant to deeper understanding of phenomena or who consider that their own way of communicating, promoting and selling cannot be improved (insofar as already perfect).

SIDE EFFECTS

In several cases, above all related to semantically complex languages ​​such as Italian, false positives have been found linked to locutions, the structure of which  “cannot not be defined no less than complex to interpret”.

The effect of the active ingredient is also undermined by the adoption of advanced expressive tools such as emoticons and emojis, especially when combined with figures of speech that make automatic interpretation of the text potentially contradictory.  Antiphrases, “using Sentiment Analysis we made a good impression!”, or euphemisms and understatements, “Sentiment Analysis sometimes provides results that are very nice to interpret”, risk diminishing or even canceling the effect of the medicine. For this reason it is advisable to administer the medicine always bearing in mind that human supervision is preferable.


Sonia Montegiove – Stefano Epifani