Complex algorithms, ontologies, and machine-learning based classifiers are needed to separate noise patterns and spam-like comments, from genuine signals. This is where the complexity arises, and the signal strength calibration algorithm has to then be customized according to each person’s behavioural patterns, lifestyle / context, and transactions.
The beauty is that you don’t have to get it right the first time, since, like any neural network, the Choice Engine will improve & learn with usage over time. Excerpts from the previous post.
3/ The Taste Graph
Even after noise filtration techniques are applied and continuously re-calibrated to distill signals from the vast external data, constructing a graph for a set of users or an individual is immensely challenging.
“Society exists only as a mental concept; in the real world there are only individuals.” ~ Oscar Wilde
Personalised Choices is the Holy Grail. The graph needs to consider each person as an individual.
Not as a segment.
Not as a class.
Not as a commodity.
It needs to respect his/her individuality, which introduces an unparalleled level of complexity where traditional analytical methods such as collaborative filtering fails, as it always tried to club people into segments / clusters.
I listen rock music, watch noir movies, like to wear tees & jeans and trek. But then I also like to eat at home, read comic books and travel in the comfort of my car. All these come together to define my personality (and I know this about myself because I know me!). You cannot recommend a book on contemporary Indian politics just because most of the people in India who like listening to “Time” or “The man who sold the world” or liked “Pulp Fiction” are young elitists and in the present day like to read about contemporary Indian politics.
Peer pressure plays a huge role in what people demonstrate about themselves in social media. But when it comes to put their money where their mouth is, they more than often buy things, which they really like.
Therefore, it is critical that the foundation of the graph is not simply built on social media affinities. The Taste Graph needs to build a persona of every individual, simulated as in the real world. This requires an element of behavioural science incorporated into the taste graph algorithm based on demographics, socio-economic context and such, scaled to millions of individuals.
Find out about the last 2 components of a successful taste graph in my next post!