After data is loaded, we enter the following dialog:
Click 【Edit】à【Model options】on the menu bar or button on the screen to go to the dialog below:
“Model evaluation criterion”: MSE (mean squared error), MAE (mean absolute error) and MAPE (mean absolute percentage error). Default is MSE.
“Percentage of test data”: Set the test data percentage.
“Set random seeds”: You can control randomness of model building through this option; default is 0.
“Number of concurrent tasks”: Set the number of concurrent tasks for data scoring on time series data.
“Timeline”: Set the timeline at data loading.
“Target series”: The variable to be scored using the data mining model, known as the scoring field. The drop-down list includes all variable names to choose. The target series must be a count type or numeric type.
“Model file”: Model file name with the extension of .tsm. It has the same default path that contains the data file from which data is loaded and has same name as the data file. YModel supports user-defined model file path and name.
“Data file”: The source data file, which can be viewed through 【View】 ->【View source data】on the menu.
“Step”: The to-be-scored nth piece of data counted from the source data in a sequential order. It is a non-zero positive integer automatically generated from 1 at row adding, or you can enter it manually.
“Time”: The future time interval predicted according to the smallest interval on the timeline.
“Score”: The result of scoring the target time series.
Click 【Execute】->【Time series scoring】or icon on the dialog to score the time series data. The scoring results will be displayed in “Score” column.
Double click the to-be-scored data or click icon on the screen to view the model performance. The blue line represents true values and the orange line represents scores.
“Score”: The scoring values of the current time series data.
“mse/mae/mape”: The criterion displayed according to the model evaluation criterion configured in “Model options”.
“XStart value”: The starting value on X axis in model performance.
“XEnd value”: The ending value on X axis in model performance. The value is later/greater than the starting value.
“Zoom”: Drag the cursor left to right or in the opposite direction to set the display size of the model performance chart.
“Keep empty values”: Set it to keep the empty values in the target sequence or not keep them.
Double click the target variable, or click 【Execution】->【Variable analysis】on the menu bar to view the variable scoring result chart, as shown below: