Smart Data Analytics Tools or Smart algorithms - Which one will give better predictive results ?
In the future, we expect all the industries & businesses to broadly apply predictive algorithms for targeted marketing, both online and offline of any thing consumed by the user on a daily basis. It can be an item or a movie or an Ad or a TV event. All industries & businesses will see a ten fold increase in Data Driven Decision making strategy in the next 3-5 years.
Recommendation algorithms provide an effective form of targeted marketing by creating a personalized experience for each customer. A good recommendation algorithm scalable over very large customer base; be it for a retailer, advertiser or broadcaster or for a business entity should cater to a wide range of products , items and to customers with diverse behavior. (eg: a teenager & an old person might use a mobile app in an entirely different way).
There is a lot of buzz around Big Data Analytics, analytics tools like Tableau , R tool, Rapid Miner, Nodexl etc. What gives better predictions ?
Where one should concentrate first ?
Should one select the tools first and then implement the prediction algorithm or one should first understand the problem to be analysed, decide on the algorithm and then zero in on the tool ?
Most of the algorithms should be smart & quick to generate dynamic predictions within a fraction of a second. It should be able to adapt to the changes in the users behavior data, and makes compelling predictions for all types of diverse users regardless of the number of users, types of users, time of usage or the location of usage.
Unlike other algorithms, item-to-item collaborative filtering is one type to meet this challenge. Another one is cohorts analysis when ever the algorithm needs to give some quick recommendation to a new user.
If there is no compelling organisation policy or push from the client to use a specific data analytics tool, I would recommend that - first analyse the problem, arrive at the predictive algorithm and then select the tool as part of the implementation strategy. Always select tools & algorithms for scale-up.
