A basic prediction model of a machine learning program works on the following steps- Acquiring Data.
This is the first step. We start by getting data we want to train our machine with. We can gather it in formats such as excel sheets, CSV, XML format etc. A lot of online resources such as Kaggle provide ready to use data sets.
Manipulating Data
Once we have our dataset, we need to make it suitable for the task. A raw dataset is not always reliable. It may have missing values, wrong order etc, which might cause bugs later on or cause our model to not be accurate enough.
Training Model
Once we have the dataset ready to use, we need to use it for training our machine learning model. However, it is always better to keep some data from being used as training set, as it can later be used for testing.
We can feed the data to the model by writing a script that will read the file, as data can’t always be fed in the IDLE. It is often very large and the best format is usually CSV (comma separated values), which is much faster to read than other formats and is usually the smallest in size.
Testing Model
Once we feed the model with the training set, we need to make sure that it is functioning as per what we want, so we test it. This is usually done with a piece of data that we haven’t fed it with. And also, using some data that is close to the values that is provided by our training set.
Improve the Model
Now that we have completed our main work, we need to focus on making our model better and better. This can be done by providing more data, and testing it more.
