Missing refers to situations in which the actual value is unknown, such as when there are unanswered items in a survey.

Missing value refers to the value used to indicate missing status.

Handling of missing values in Easy Predictive Analytics

When items used for learning are missing

Easy Predictive Analytics creates a prediction model by giving the missing data a “missing” status. So, for example, sales data for a product is treated differently if sales are recorded as “0 yen” than if sales are not known and "” (it is blank because the value is unknown). In general, a better prediction model can be created if there is less information missing for items related to the item you want to predict.

When the item you want to predict is missing

If some of the items you want to predict are missing, Easy Predictive Analytics removes the missing data and learns and evaluates the prediction model. Keep in mind that if the item you want to predict contains too many missing values, you may not be able to create a prediction model.