How Dataset is needed for Machine Learning Model
Data is the basic need of any machine where we talk about developing it into an Artificial Intelligence type of device. A programmer has to have a plan of what they want the machine to perform and has to feed data of only that kind, but as machine learning models are intelligent enough with the programs they are built-in, they can learn as they perform tasks.
Siraj Raval speaks about what kind of data is to be selected, how to process it and then transform it to be fed in the machine learning model. The data which he refers to here is the brain scans of humans to classify if whether or not someone is meditating.
Siraj shows how we can use Python language to extract data from brain scans and feed that dataset into single layer known network created using TensorFlow. To know if a person is mediating or not, the network will create a line between two classes. Data can come in the form of Text file, Relational Database or CSV.
To understand the process of creating a dataset using python, please check out this video.