.png)
Smart predict can be used to create extended predictive scenarios, which can become very complex.
In SAC, we can have 3 different smart predict scenarios.
1. Classification Yes/No predictions
2. Regression predict values based on projected charges
3. Time series predict projected values based on time periods
Classification
Classification is used to rank a population relative to the probability that an event will happen.
Smart Predict Classification is useful to answer questions of this form:
“Who is likely to <event> next <period length>?”
for any systems to predict scenarios. the system should be trained with sufficient data. The quality of data is vital to build a reliable predictive model.
here we are taking a data set of passengers who traveled on the titanic ship.
first, we will train the system with a dataset with full details of passengers.
then we take a new data test which doesn't have much information about passengers. with the help of Machine learning. our system will be able to predict the survivors of the titanic accident.
we need to create datasets before building classifiers.
lets create datasets (train & test)
now lets upload the test dataset.
now we have the datasets ready. lets see how we can train our system for smart predictions.
building classifier
select the datasource to train the system.
we can edit the columns in the dataset
as passenger id is the unique identifier, I am setting it as Key.
our intention is to train the system about the survivors in the titanic accident. lets set the target as survived. click save to train the system with this data.
now our system is trained and we have the influencers and performance indicators.
we also know what will be the percentage accuracy of prediction.
now we have our system trained with data.. lets apply this prediction scenario to a test dataset.
click on the apply predictive model
select our test dataset as a data source.
and also select the columns needed in the output file.
we can see the status of the prediction at the bottom of the screen
lets see the result file to know the output of the prediction
we have an output with a prediction category.
in which 0 indicates passenger dint survive and 1 indicates survived.
Comments
Post a Comment