spatio-temporal and geographically weighted regression analysis
Feature Description
Spatio-temporal geographically weighted regression is an extended and improved version of geographically weighted regression, which can analyze spatial coordinate points with time attributes, addressing the overall spatio-temporal non-stationarity of the model.
Application Scenarios
- Study the changing trends of urban housing in terms of time and space.
- Study the factors of provincial economic development and their spatio-temporal patterns.
Parameter description
| Parameter Name | Default Value | parameter interpretation | parameter type |
|---|---|---|---|
| Source Data | Specifies the vector dataset to be calculated. It can be point, line, or region dataset | DatasetVector | |
| explanatory fields: | Explain the field name. Explanatory variables are independent variables, i.e., X in the regression equation, used to model or predict the values of dependent variables. | String | |
| kernel function type | GAUSSIAN | kernel function type. | KernelFunction |
| modeling fields | Modeling fields. Dependent variables, i.e., the variables to be studied or predicted, only support numerical fields. | String | |
| time field | Set the field representing time. | String | |
| Time Unit | 天 | Set the time interval unit. The time distance is the difference between two records in the time field. The time distance unit needs to be uniformly converted to the specified time distance unit, supporting Seconds, Minutes, Hours, Days, Weeks, Months, Years. For example, if the time interval is 60 minutes, the corresponding time distance is: Minutes: 60, Hours: 1, Days: 1/24, etc. That is, the unit can be selected as needed. The smaller the time interval unit, the larger the time interval distance. Depending on the set kernel function, it reflects in the weights, making the weights larger or smaller. | TimeSpanUnits |
| Bandwidth Method | AICC | bandwidth determination method. | BandWidthType |
| bandwidth type | FIXED | Bandwidth type, providing two types: fixed bandwidth and variable bandwidth. | KernelType |
| spatio-temporal geographically weighted regression prediction dataset (Optional) |
spatio-temporal geographically weighted regression prediction dataset | DatasetVector | |
| the specified datasource to save the forecast result dataset (Optional) |
the specified datasource to save the forecast result dataset | Datasource | |
| specified forecast result dataset name (Optional) |
specified forecast result dataset name | String | |
| field mapping for forecast data (Optional) |
Field mapping for forecast data, indicating the correspondence between the fields of the prediction dataset and the fields of the source dataset. If not set, the prediction dataset must contain all explanatory fields from the source dataset. In WebGPA, the filling method is {"inputFieldName":"ABC","predictFieldName":"DEF"} | Object | |
| Result Datasource | Specifies the datasource where the result dataset is stored | Datasource | |
| Result Dataset | Specifies the resulting dataset name | String |
Output Result
| Parameter Name | parameter interpretation | parameter type |
|---|---|---|
| results of spatio-temporal geographically weighted regression analysis | Analyst result includes geographically weighted regression summary results and result dataset. The interpretation of result parameters is consistent with geographically weighted regression analysis. | GWRAnalystResult |