The future of technology is Sentiment Analysis (SA)

With ever increasing human-machine interfaces, our dependency on social media, more sensors embedded with internet of things, personal assistants or other robots, sentiment analysis is one of the most exciting challenge for computerized services to understand, and potentially predict, anything to do with the human being complexity.

Sentiment Analysis and social media

While using social media platforms, I can express my opinion and share my sentiment with a large public or limited community. Whether you like my idea, my post, my picture or my video, it's one click away for you to express a positive or negative feedback - please feel free!

The average time spend per day on social media is increasing as it represents the main stream to stay in touch with family and friends. Based on recent survey made in the US, about 80% of connected Americans are using Facebook and 30% LinkedIn (see below).

The human language is quite complex for machine based learning systems to interpret opinions being potentially reflecting sarcasm or irony, which goes far beyond applying a predefined set of words (lexicon) and provide some sort of weighting to a sentence; this is primarily the reason why tweets and other posts are not immediately deleted when breaching terms and conditions, requiring someone to report.

Social media platforms are able to suggest ideas about who should I connect to and perform sophisticated analytics for promoting targeted products or advertising campaigns but none can't predict my mood today or what I am looking for at this particular point in time...

The future for social media is to provide more customized services, predict my sentiments and opinions while increasing overall productivity.

Sentiment Analysis to support my assistants

Over recent years, we have seen some significant investments from major market players to expand the field of personal services and assistants - such as Siri from Apple, Cortana from Microsoft and Now from Google - through voice-enabled speaker that helps to search things with natural language queries.

- chabots are bringing some forms of humanity to IT but will not replace human feelings - unless sentiment analysis will be able to combine facts with sentiments; in other words, it is hard for a chatbot assistant to manage customer's anger...

- Same for smart assistants: they are able to answer almost any questions, predict time to leave for my next appointment depending on traffic jam and many other things; unfortunately, we (almost) feel frustrated that it correlates only known events - eg. your flight departure has been delayed, please leave later - rather than predicting what to do to optimize your time and increase your productivity - eg. today, you'd rather focus on this task to optimize your efficiency.

Beside strategic investments, the level of adoption remains marginal in the consumer market and I haven't meet, so far, anyone considering his/her smart assistant as equivalent and good as a personal assistant.

Sentiment Analysis and security

In everyday's life, we see a massive uptake for sensors and payments but not emotion.

All devices enable people to pay - from banking cards, smartphones or watches - with many authentication protocols (fingerprint, face recognition) but none with any form or shape of sentiment analysis - eg. am I buying under stress condition, high blood pressure or being in hurry?

In the cyber-security field, leading vendors are developing big data analytics to prevent transactions or processes to happen based on potential anomalies. It is particularly true for network-based detection which, after building historical database of events to define standard patterns, can monitor the traffic in real time and correlate every single transaction with this database to authorize or reject it.

In the forefront of security, behavior monitoring and sentiment analysis are key success factors to provide advanced analytic and help to identify complex business scenario as cyber-attacks get more sophisticated with social engineering.

Sentiment Analysis and technology disruption

For many businesses, significant technology disruption is expected to come from changes such as artificial intelligence, machine learning or robot advisors.

One of the driver for sentiment analysis is the early detection of business opportunities and warning signals; from commercial management to financial investment strategy, the field of application for early detection of signals, whether positive or negative, to influence any human decision making process is huge. It could facilitate to mitigate negative signals and anticipate positive signals earlier, definitely a win-win.

If we could combine sentiment analysis opportunity to detect signals with technology disruption, it could bring a significant competitive advantage to many industries and particularly for early adopters. Market is about speed and quantity of information available, meaning (almost) impossible to handle for any human being.

There is a vast range of opportunities for Sentiment Analysis in technology - almost as diversified as human complexity is!

CR