How supervised and unsupervised classification algorithms work

In this video I distinguish the two classical approaches for classification algorithms, the supervised and the unsupervised methods.

We provide one visual exemple on how both strategies works and suggest examples of algorithms of both types.

Videos about the mentioned supervised algorithms:

Videos about the unsupervised algorithms:

[samuelmungu]

thanks for the video.. it really been so helpful

[danielvalentins]

Sotaque inconfundível mesmo!
Muito obrigado por esse vídeo, meu TCC usará classificação e esse auxílio visual deixou mais tranquilo o entendimento. Já me inscrevi e irei divulgar o canal, abraço!

[catalyst]

So to get this straight, with unsupervised classification algorithms we can visualize how similar the data points of various objects are without knowledge as to what their actual attributes values are in order to group them. With unsupervised classification we never learn how instances are grouped? we just visually perceive similarities of the features of the points and group them accordingly such as by using K-means. Whereas, supervised classification algorithms tend to learn how data points are classified accordingly (training set) by looking at their attribute values before actually building the model (test set) with unknown instances in which we look into their attribute values to group them accordingly. Correct me if I am wrong please?


Oh and also just to add to what you said in the video, with unsupervised classification we have to validate whether our algorithm has given us a valid result. Whereas, this isn't a necessary step with supervised classification techniques since the data has been studied to classify instances accordingly.

[Ivsucram]

HI Thales, I have a question.

Can I use a pure unsupervised learning, without supervised learning process at the end, to classify a dataset or I always use the unsupervised learning just to detect features and "pre-train" or "initializate the weights" of a supervised learning?

BTW, I would like to make a request. I saw that you have a video for SOM (that actually helped me a lot). Can you also make an video about ASSOM? Your videos really help me to the introduction of the theme.

Thanks. 

[iot-nlp-daily]

please share some lectures on active learning and semi-supervised learning. Thanks.