I’m currently working on a Masters in Data Science, where I’m seeing some interesting bits of information that I’m going to start sharing here. This is my first post in this series.
Dictionary.com defines the verb “predict” as “to declare or tell in advance; prophecy; foretell.” To me, this is the traditional connotation of the word and what many people think of when they hear data scientists talk about “predictive modelling” or “predictive analytics.”
Webster’s dictionary defines the verb predict as “to declare or indicate in advance especially : foretell on the basis of observation, experience, or scientific reason”. The last part of this definition is closer to what we are talking about when we use the work “predictive” in a data science context. We are trying to “estimate an unknown value” (Provost & Fawcett, 2013). In order to estimate that unknown, we look at data where the value is known and, depending on the type of data and what our “target” is, use various processes to segment the data to the point where we can say something like, “cases with the attributes a, b, and c are more (or less) likely to x.” This is not a “prophecy”, this is a prediction that is based on both observation (existing data) and scientific reason (the process to segment the data.)
Webster’s Dictionary: https://www.merriam-webster.com/dictionary/predicting
Provost, Foster & Fawcett, Tom. 2013. Data Science for Business. O’Reilly Media.