The bar chart at 5:22 is not showing colts and fillies have a higher chance of winning. It is merely showing that there is a higher population of colts and fillies than geldings. You need to compare win rates against the population of each gender.
ML and data science is so exiting! I wish I had more time to get more into it. What are you goals with machine learning and data science? Do you learn it just for fun or do you see real potential for you business?
Wouldn't a regression here not better suited? A classification model gets less feedback in training, because when it predicts place 2 for the winner, it is as bad, as place 6 for the winner (for a classification). But indeed, predicting place 2 for the winner is a much better prediction than place 6. And a regression, probably would then just predict something like place 1.78 instead, so you know this horse usually will be 2nd, but might as well win, too by a close similar chance and I'd bet the even if you just round the predicted results to the next integer, the output will be more precise than the classification training as for the learning algorithm there is more feedback.
1:26 - it is a production database, however its an OLAP database not an OLTP database so for fast queries that you're used to with RDBMS you won't get instant results. for inferences however with BQ ML you can use it for front end apps
Why did you not create the dataset of actual position = 1. Also on the predicted set. What is the key data to be used against the set you trained on. For example the racecard does not give you the same set of features to compare.
How can you apply this to real data? you would need the data in the proper format, but who provides this? If you have no real data, you have no logical conclusions...
Be sure to read honest and real reviews of The Horse Race Predictor on my blog before you buy. Go to gohonestreviews. com/the-horse-race-predictor-review/ Thanks, Osborn.
This was great, thanks, I have been researching "top horse racing tips today" for a while now, and I think this has helped. Have you heard people talk about - Honackson Meyrayden Idea - (Have a quick look on google cant remember the place now ) ? Ive heard some super things about it and my partner got cool results with it.
The bar chart at 5:22 is not showing colts and fillies have a higher chance of winning. It is merely showing that there is a higher population of colts and fillies than geldings. You need to compare win rates against the population of each gender.
need more videos on ML stuff. 😍
Back to the Future 2! So cool!
ML and data science is so exiting! I wish I had more time to get more into it. What are you goals with machine learning and data science? Do you learn it just for fun or do you see real potential for you business?
I used to be really into ML on Kaggle. I see potential for certain types of businesses, but you really need to have access to high quality data.
Wow, I never knew Google had such tools. Thanks for showing them!
How do you make the animation on your videos ?
This guy has never bet on horses
Wouldn't a regression here not better suited?
A classification model gets less feedback in training, because when it predicts place 2 for the winner, it is as bad, as place 6 for the winner (for a classification). But indeed, predicting place 2 for the winner is a much better prediction than place 6.
And a regression, probably would then just predict something like place 1.78 instead, so you know this horse usually will be 2nd, but might as well win, too by a close similar chance and I'd bet the even if you just round the predicted results to the next integer, the output will be more precise than the classification training as for the learning algorithm there is more feedback.
Now I understand ML. Better than any Google video.
cant regret subscribing to this channel....brilliant stuff
Nice one super excited to get started with this tutorial.Thanks for this
1:26 - it is a production database, however its an OLAP database not an OLTP database so for fast queries that you're used to with RDBMS you won't get instant results. for inferences however with BQ ML you can use it for front end apps
Seems like RUclips thinks I’m a bot but I’m not :(
Jeff is my favourite tech youtuber!
Why did you not create the dataset of actual position = 1. Also on the predicted set. What is the key data to be used against the set you trained on. For example the racecard does not give you the same set of features to compare.
I just love you fireship.😘
Love you too
sus@@Fireship
Thank u for this awesome video
Definitely need more ML stuff !!!
So cool!!! best ML explanation! Need more!!!!)
turn to speed .75x its exactly wanted to hear
Amazing video quality!
right to the point - thanks!
More of this please 😍
you're a good teacher. We can't practice. Horse data link is dead (404)
Yessssss. Jeff this is GOLD!
Unfortunately the horses.csv file has disappeared off Kaggle
Just remember that if you using the cloud you are practically doing the work for other people
what do you mean? explain.
What did you mean?
If seems the dataset is no longer available on Kaggle, do you know where we can download it?
This doesn't work for me, the model won't create the ML. Has multiple errors when trying to create itself
Anyone who can offer up help on this?
Appreciate the video but please slow down for us non computer guys.
Build a ML model without taking care of overfitting is a big no. Could have created a 4 fold CV with L1 or L2 regularisation.
Awesome ❤️
I cannot believe only 27k watching
any new update with tensorflow?
fireship thanks
Nice
Cool.
How can you apply this to real data? you would need the data in the proper format, but who provides this? If you have no real data, you have no logical conclusions...
Indeed. I have the data
So where is the results? How much money did you make? o/
is this David Aragona?
it's his boyfriend
Be sure to read honest and real reviews of The Horse Race Predictor on my blog before you buy. Go to gohonestreviews. com/the-horse-race-predictor-review/ Thanks, Osborn.
👍
This is useless in terms of actually learning something about horse races prediction, you basically imported a model that we know nothing about.
Do you have to pay to use BigQuery?
Yes, but it has a serverless pricing model, so you pay for what you use. It's cheap compared to similar alternatives.
This was great, thanks, I have been researching "top horse racing tips today" for a while now, and I think this has helped. Have you heard people talk about - Honackson Meyrayden Idea - (Have a quick look on google cant remember the place now ) ? Ive heard some super things about it and my partner got cool results with it.
2nd
you should talk even faster
First :P
Love the linear_regression dog gif 😍
@@darshangowda309 Haha, it was hard to find the right linear regression gif
2nd