To conduct predictive analytics in Easy Predictive Analytics, a dataset(training data and prediction data) is required. This document explains how to create datasets.

*A large sample dataset is included with Easy Predictive Analytics. See the tutorial for interesting examples. See here for information on how to open a tutorial.

Predictive analytics is an analysis method that uses AI technology to predict future results based on past performance data.
Predictive analytics is like a weather forecast in business.
Predictive analytics enables faster and more efficient decision-making in many aspects of your business.

It has a wide range of applications and brings efficiency to a wide range of operations including marketing, production management, customer support and assessment.

Predictive analytics involves two steps: learning and predicting. The following is an example of predicting a customer’s contract from actual contract data.

Prepare data with the variables of input information and the variables that represent the actual results you want to predict. The following is an example of a close prediction.

Easy Predictive Analytics can handle dates, strings, numbers, and text. Any necessary preprocessing is done automatically, such as leaving the unknown blank. If you would like to see how datasets are used inside Easy Predictive Analytics, see “What is a dataset?”.

Prepare the dataset as a csv file (comma-delimited text file) or as a tsv file (tab-delimited text file).

If you would like to learn more about topics such as the timestamp formats that can be used in datasets, see Datasets > Data and time format”.

In predictive analytics, the accuracy of predictions is determined by the scale and quality of the prediction model creation (training) data used to create the prediction model.

• The more rows, the more accurate the prediction. It needs at least 100.
• The more variables, the more accurate the prediction. You need to prepare variables related to the variable you want to predict. You should not include variables that are not clearly related to the variable you want to predict.
• You don’t know how accurate your dataset is until you actually learn. Easy Predictive Analytics makes learning easy, so try it when you’re ready.

So far, we have discussed the prediction model creation (training) data required for learning (prediction model creation).
When you create a prediction model and use it to obtain prediction results, you prepare data for prediction.
Prepare the same input variables as the prediction model creation (training) data for the target you want to predict. You don’t need the variable you want to predict.

In this example, the close probability is predicted for the customer in the prediction data.