Data Analysis Mechanism

The mechanism acts as a combination of objects in the 1C:Enterprise script. The interaction flow for the main mechanism objects is demonstrated in the figure.

The objects that can be used by the 1C:Enterprise server are shown in the white boxes while the objects that can only be used by the client are shown in the gray boxes.

Customizing data analysis columns – a set of settings of source columns for data analysis. These settings specify the data type and function of each column, as well as additional settings that depend on analysis type.

Data analysis parameters – a set of parameters of the executed data analysis. The assortment of parameters depends on the analysis type. For example, for cluster analysis, you need to specify the number of clusters to split the source objects into, the type of object distance dimension, etc.

Source data – source of data for analysis. You can use a query result, an area of cells in a spreadsheet document or a table of values as the data source.

Analyzer – an object that is directly responsible for data analysis. Specify a data source and parameters for this object. The output from this object is a data analysis result with its type depending on the type of analysis.
Data analysis result is a special object that contains information about the analysis result. A different result is designated for each type of analysis. For example, an object of the DataAnalysisDecisionTreeResult type will be the result of an analysis of a decision tree. You can then export the result into a spreadsheet document, using the data analysis report builder or using program contents that give access to results. You can also use the result to create a forecast model. Any data analysis result can be saved for subsequent use.

Prediction model – a special object which makes it possible to create a prediction based on source data. The model type depends on the data analysis type. For example, the data analysis model for an association search will be of the PredictionModelAssociationRules type. This type of model can provide predictions along the lines of, since this customer has purchased certain goods, there is a certain probability that this customer will also purchase other goods. For input, the predictive model receives source data to be used in the prediction. The output is a table of values containing the predicted values.
Data source – a table of values, query result or a spreadsheet area that contains information to be used to create a prediction. For example, for the rules of association prediction model, the selection may contain a list of goods from a sales document. The result we get when using the model may suggest which other goods we might offer to the customer.

Selection column setting – a set of special objects that define the mapping between the forecast model columns and the forecast selection columns. For example, you can map Goods forecast model columns to the Nomenclature selection column.

Result column setting — specifies the columns to be placed into the resulting forecast model table. For example, for an association search, we can export the nomenclature of what the customer is likely to purchase and the probability of this purchase to the results.

Model result – a table of values that consists of the columns specified in the resulting column configuration and containing the forecasted data. Specific content is determined by analysis type.
Data analysis report builder – an object that allows to output a report of data analysis result. Report builder also provides special objects for data linking. The user can use these objects for interactive management of analysis parameters, data source column settings, forecast model column settings, etc.

Analysis Types

This mechanism can be used for the following analysis types:

• General statistics
• Association Rules
• Search for sequences
• Decision tree
• Cluster analysis