IBM Watson: Using the Natural Language Classifier in IBM Bluemix for IMDb Movie Ranking - Part 2!

 This time we're gonna be going over part two of my natural language classifier service tutorial for bluemix and i just released this a few days ago that was a part one now we're gonna be doing part two in part one I showed you how to start the training process of your classifier and now after that I decided I'm going to make a part two once it's done training because of course it's a very large training file so if you haven't already seen part one please do there will be a link in the description and it'll also probably be in the related video somewhere however let's get to the point now okay so now what we're going to be doing is I'm telling you first of all it took this I'm not exactly sure how long because I was sleeping but it took this around six to eight hours to train so if you ever set it to Train know that it's going to take that much time so I'd recommend doing it at night when you're about to sleep also this doesn't always get it that right due to the fact that it can only have 15,000 movies in its database they may seem like a lot to a human but to a computer that's a very very little amount so now fifteen thousand movies gets the job done but some things are gonna be slightly off and again that doesn't get it right to the decimal but anyway now I'm I'm assuming that you've checked your natural language classifier the status of it I believe you have it's done training for you and so now I'm going to be teaching you how you can actually run classifying requests on it so first of all let's move over to the natural language classifier service documentation again and then in the documentation what we're gonna be doing is under here over here click on get getting started and then click on classify text okay then you see this big command here copy that entire thing paste it into sublime text or your favorite exeter then copy your username or actually just your username your password from another command paste it into this one perfect copier classifier ID and paste it in to the URL over here and then instead of how hot will it be today type in something like or actually a movies name so I'm just going to go to IMDB and see what we can do okay so now let's take a movie like poem alone first let's do this ok so we see that the rating on IMDB is 7.4 and so if I come over here and I type in home alone I copy this command and I paste it into my terminal oh now again before I continue before I click enter home alone is not in the movie database I repeat is not in the movie database this is due to the fact that I was only get for able to get movies from numbers to see due to the fact that it only allows 15 thousand movies and so home alone isn't even in the data set this is using machine learning to find patterns between all the movie names and their ratings to see what home alone will be rated so now if we click enter it gives us 8 which is quite close but it's point 6 off let's go for something a bit more uh knob vyas for the computer or actually obvious so something like oh yeah this is a good movie the towering inferno listen ok so as we can see the towering inferno is rated 6 point 9 on IMDB so if I copy this movie title and instead of home alone if we type in the towering inferno as you can see it gives us 7 with 96% confidence that's great before for home alone it was only 67% sure now it's 96% sure for the towering inferno so now as you can see the rating is only point one off so it's extremely accurate in that sense because of course machine learning is a powerful tool next let's go for something like it's a good movie I know the name of at least From Here to Eternity I guess okay so now let's copy this one and let's paste this one in instead of the towering Auburn and for now it's about paste oh that doesn't work okay so I guess we're gonna have to edit the command in sublime text oh yeah that's probably why okay so I'm going to type in From Here to Eternity copy it and paste it into my terminal enter and it gives us an eight with 97% confidence and it is actually seven point eight so that's a point two difference next let's get Titanic that's a classic let's copy this or you know what actually I'm just gonna type it in so if I clear and now if I get so it's an ik see what it gives us seven whereas it's actually seven point seven point seven off it's okay now one last this is gonna be the last one just so you can really get the point of it frozen which was just released a few years ago so frozen and Watson says eight whereas it's actually seven point six so it's point four off so as you can see on average watching Watson only gets less than one rating off and so this would actually be a really helpful tool for two purposes first of all a consumers perspective and also for movie makers themselves if we were to actually train this on many many many more movies or even maybe the entire IMDB database in general what we could do is first of all let's say a consumer really wants to see what can i watch right now that I'd like well all the users need to do is enter his preferences his ratings to some movies that he's watched before it'll compare it with the IMDB preferences of millions of other users and then give you a rating of some movie that it sees in the IMDB database that it thinks you'd like that would be great and so that's one thing where it could help consumers but the movie makers how could movie makers we benefited from this simple it's like they just enter in the title that movie may be a description or something and they'll be able to see how much their movie may make or they'll be able to decide the budget or something like that and they'll very early on be able to see what the rating of their newly released movie is without even having to release any votes polls or anything just use machine learning okay now that was actually gonna be it for this tutorial again if you liked please leave a like down below you can also comment if you have any questions suggestions concerns pretty much everything else you can also email me at IG manage email comm if you have a video question that you'd like to get featured in one of my next videos you can also contact me at tad you Manny on Twitter and that's going to be it sorry subscribe to my channel if you're new or you like my cotton you want to see more of it so that's going to be it for this tutorial goodbye