Big Data critical to semantic search and machine learning

How Semantic search Works: Lesson #3 - Data

We live in the age of Big Data to such an extent that we take terms like “data mining”, “sentiment mining”, “metadata” and “data analysis” for granted. We barely stop to think that just ten years ago a lot of the things we now assume are normal, were simply not possible. 

Machine Learning and by extension semantic search are direct expression of the the Big Data problem. Data is not just what we are now capable of collating in a larger volume than at any other time in history but also what we each generate. My weather report request this morning, on my smartphone, generated data. It gave away location, it was added to my profile fine-tuning the ‘signature’ of my behavior. It became part of a wider pattern of activities that took into account who I am and what I do, the device and place and time of day and time of year and it added one more tiny layer of data to a global profile that is then used in predictive models. 

That one ‘simple’ thing I did which tool all of 15 seconds probably generated more data and derivative signals than a whole week’s work in 1995. The reality of 21st century life is that we swim in data, we are surrounded by it. Frequently we are defined by it. 

Semantic search became imperative because the Boolean search model that relied on statistical analysis of probabilities was no longer good enough to cope with the volume of data being put on the web each minute of each hour of each day. Something better was needed in order to index it properly and make sense of it. 

At the same time, in a chicken-or-the-egg kind of situation semantic search and machine learning cannot take place without massive volumes of data being available to train them in the first place. This is particularly true of any machine learning algorithm. 

Since Data is the metaphorical road search and machine learning run on and data is also the fuel that drives them and helps them develop we have a clear path to cutting through the complexities of their make up. 

It’s no secret that Big Data labors under the vectors collectively called the 4Vs that both define it and challenge it: 

  • Volume
  • Velocity
  • Variety
  • Veracity

The last of the four, Veracity, is the ultimate key that truly unlocks the value of the other three. 

Trustworthiness is Key

Veracity cuts both ways. Those who trawl for data need a way to verify its integrity before they start to process it. Those who use it (as in using search to find answers, for example) need to have some degree of confidence in what they are presented with. 

That also becomes the key we need to unlock the secret of success with search and visibility. Whether in search or social media platforms or even in apps that find and serve content, in order for any business to appear, they need to have a credible, detailed, data-dense presence on the web that inspires loyalty and trust. 

The only way to achieve that is if you have the kind of marketing strategy in place that, in the mind of your audience, answers these questions in a definitive way:

  • Who you are
  • Why you are
  • Why you matter to them

It might sound simple but it isn’t. Solve it however and most other things begin to take care of themselves.  

Get smart: SEO Help: 20 Semantic Search Steps that Will Help Your Business Grow is a practical step-by-step guide to applying semantic search principles to your business.

Additional Resources

How semantic search works – Lesson #1: Classification 
How semantic search works - Lesson #2: Validation

   

© 2017 David Amerland. All rights reserved