-
Machine learning can be useful to recommend movie at netflix
-
or how many homerun that baseball player can hit over the course of his career
Why Machine Learning useful:
-
Have some great journey along statistic
-
Grew out of computer science, many fields use it for making predictions based on data
-
predictions based on classifier, recommendation systems among the many
-
Statistic used based on existing data,drawing conclusion like for example(difference left and right batting)
-
Machine Learning focused on making predictions. It doesn't care about the data model, as long as it making right prediction (how many homerun)
-
Supervised learning learned from label data. Labeled(spam or not spam) or given parameter's of house predict (labeled house prices)
-
-
For example trying to separate the photos without any structure of data.
Kurt's in Twitter(social data) just focusing on the techiniques involving these data
-
Clustering (K-means, hierarchy)
-
Dimensionality reduction(PCA)
-
Often combination between the two(to understand better of cluster, using pca to reduce the dimensionality)
-
These examples is given bunch of infos predict how many homerun the players would have
-
To do this, we have some existing data (infos and homerun) to make a model
-
Feed the existing into a model and then making the prediction
