R Stats: Naive Bayes and k-NN

Подписаться 2 тыс.

Просмотров 27 тыс.

50% 1

This video shows how to use R to construct and improve a Naive Bays classifier. It compares the performance of Naive Bayes classifiers against the popular k-NN. Finally, the video explains the need for approaches better than mere splitting of the data into training and validation sub-sets. The explanation will be quite informal and will avoid more complex statistical concepts.
The data for this lesson can be obtained from (note different file name):
* jacobcybulski.c...
The R source code for this video can be found here (some small discrepancies are possible):
* jacobcybulski.c...
Videos in data analytics and data visualization by Jacob Cybulski, jacobcybulski.com.

Опубликовано:

11 сен 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 18

@ironfrown 7 лет назад

Lalith and Wei, specially for you I have added the link to my sources for the video :) I normally do not do this as the best way of learning R is to actually type it in as watching. I may start adding these in the future for my other videos.

@fathialwosaibi4024 8 лет назад

Thank you very much for taking the time to do this.

@vamsikrishna3205 7 лет назад

Thank youuuuuuuu so much ........ you have solved my 3months problems in just 15minutes...............

@raimohaikari707 7 лет назад

If you are new to R & NB classifier, you should definitely check this tutorial. It's quite informative and easy to follow.

@SabaElahi 8 лет назад

Thank you, it was extremely useful

@joesaddition8413 3 года назад

Could you please tell me about LGA? How it becomes 'factor' and why my PREDS doesn't come with any value?

@weiqi5747 7 лет назад

Can we download your scripts somewhere? Really love the videos and would like to practice more!

@timscongack5867 6 лет назад

Question on getting an error: > table(preds,d.validation.sample$vol_leave) Error in table(preds, d.validation.sample$vol_leave) : all arguments must have the same length - what might be causing this aside from incorrectly named variables?

@ironfrown 6 лет назад

Tim Scongack, this happens when the model you are using hates missing values and automatically rejects all observations with missing values, in which case the vector of predictions may be of different length than the original data. In which case, make sure you deal with missing values in your validation set as well - this may be tricky as you may lose lots of data when deleting observations with missing values, or else you need to ensure same scaling factors are used in training and validation, and ensure you are not introducing bias in validation in the process.

@TheLoggic 6 лет назад

Geat video! Cheers! Re u Australian?

@ironfrown 6 лет назад

Yes Max, Aussie, Aussie, Aussie, Oi, Oi, Oi. Living in Melbourne but the accent is part Polish :)

@SchreiEs 2 года назад

Has anyone told you that you look like Fred Armisen?

@ironfrown 2 года назад

Hahaha, nobody told me this - may be some likeness but I'd have to lose a few pounds to be a mirror image of him 😁

@ironfrown 2 года назад

Although some people in the past suggested I look like Mike Myers (Austin Powers) and others thought I look like Marvin Minsky (the grandfather of AI), may be all depends on the hair style I fashion at the time 😁