My college roommate had this posted on his bedroom door:
You can’t. They’re too fast.
In a nutshell, that ambiguity illustrates the difficulty of what is termed Natural Language Processing.
Question: How do you get down from an elephant?
Answer: You don’t. You get down from a duck.
Only 23 shopping days until Christmas!
Can we predict what you will buy, based on your previous behavior?
Should we send you an ad for the next Lisbeth Salander novel? Will you buy if we do? Will you buy if we don’t? In order to figure that out, we will be attempting to perform some Uplift Modeling.
We can segment our audience into 4 groups:
- Leave Me Alone: I’ll buy, but only if you’re not hectoring me with advertising.
- Sure Thing: I’ll buy, regardless of your advertising.
- I’m Not Listening: I won’t buy. Period.
- Frasier: I’ll buy, only if you make me an offer.
Q: Can PA of these blog posts be used to predict the content of tomorrow’s blog post?
Stay tuned to see the answer.
Again with the mysteries!
Can anyone guess what the topic of this post is, or will be once it is written?
Can a machine learn from the mistakes of others? Can it at least learn from its own mistakes?
The Credit Score is a classic example of predictive analytics. A pile of information about you and your financial history is mathematically combined into a single number, your “credit score”, that purports to predict your future behavior. Paying your debts, etc.
See the website smalltime.com for an amusing illustration of a decision tree, wherein you assume the guise of a dictator or a sitcom character, and the computer attempts to determine who you are.
Over the years, my neighborhood has experienced ebbs and flows of trick-or-treaters. Some years, no scary monsters or fairy princesses come to the door, and we end up eating all the candy in the bowl ourselves. Other years, we get so many visitors that we run out of treats, and then turn off the lights to pretend we aren’t home.
What’s the difference?
Based on our very informal observations, it is not the weather: rainy or fair, warm or cold, it doesn’t seem to matter. And that seems, strange, because our intuition is that trick-or-treaters would be very sensitive to that, and would find inclement weather to be very discouraging.
wherein the author regales us with tales of emotion, and stock markets.
Question: How can a collection of blog posts on LiveJournal be used to forecast movements of the stock market?