The State of Sentiment Analysis

 |  Posted: by
The state of Sentiment Analysis

 

The state of sentiment analysis is disappointing. After the initial enthusiasm about the concept and the flood of many tools trying to measure it, companies are searching for better ways to measure and apply sentiment analysis to business problems.

 

This was apparent at 2015 Social Media & Web Analytics conference, organized by Innovation Enterprise in San Francisco. Many attendees came from big tech companies such as LinkedIn, Twitter, and Yahoo. I heard a clear consensus about how problematic and inaccurate Sentiment Analysis was when left to machines counting positive and negatives comments.

 

The Challenge of the Human Language

 

The human language is too complex to leave it solely to algorithms. Emerging terms, negation, sarcasm, word roots (stemming), sentence completeness, term frequency, semantic ambiguity, lack of context, and differences in topic categories, among others, are issues undermining the accuracy of the Sentiment metrics currently used. Consequently, we need humans to verify and correct the models we use to support machine learning.

 

Some users of Sentiment Analysis try to go beyond simple counts of positive, negative, and neutral comments by:

  • Measuring how strong the sentiment is using rating scales
  • Identifying human emotions expressed in those comments

 

The Text Analytics Tool Race

 

Nonetheless, we still need algorithms to analyze the large amounts of unstructured data generated by social media and customer feedback systems, among other sources.

 

As a result, the data scientist community is still searching for the best ones. More text mining tools (e.g. OdinAnswers, Bitext, Monkeylearn, Luminoso, etc.) are coming to the text analytics space. However, some companies have decided to develop their own proprietary tools. It seems they have yet to find one that satisfies all their needs or that is affordable.

 

Metrics Beyond Sentiment Analysis

 

Above all, aware of the limitations of this metric, some companies use Sentiment as one more variable, among many, in their analytics arsenal. They combined it with others in the context of the business problems they face to:

  • Increase customer retention
  • Resolve customer experience pain points
  • Identify what customers like
  • Optimize customer service by matching customer service representatives with customer issues
  • Optimize pricing
  • Measure social media ROI

 

In conclusion, my hope is that the frustration with the current state of Sentiment Analysis keeps driving the development of better tools.  We need text analytics tools to extract insights from unstructured data faster and in a more cost-effective way. In the hands of market researchers, these tools can facilitate and speed up the analysis of qualitative data and open-ended survey questions.

Only logged in users can leave comments.