Detailed Steps

This section will provide a detailed introduction to the workflow of PrCore through a hypothetical use case, as well as an introduction to the main features of PrCore.

Assume that you already have a business event log file in a certain domain, which should only contain completed cases. Your business event log can come from your business management system or other data sources. The file formats supported are CSV, XES, and ZIP. PrCore can train a model using this log file to provide predictive results and various prescriptive suggestions for your ongoing cases.

First, you need to use the method described in Upload File to upload this file to PrCore. Then, use the API described in Provide Column Definitions to tell PrCore the meaning of each column in your log file. Finally, you can use the method described in Create Project to tell PrCore which cases are successful and as expected, and which actions and events can be attempted as interventions for improving a case based on your business domain. PrCore will create a project based on your event log file and this information, and then automatically start training the model. After creating the project, you can change its name and description at any time using the API for updating basic information, so that you can better differentiate between different projects. The API can be found at Update Basic Information.

PrCore can be connected to multiple plugins, each of which is an independent container responsible for providing a certain type of prediction or recommendation. Therefore, when creating a project, you can specify the parameters of the plugin according to the instructions in Set Parameters, which will adjust the effect of the trained model of the plugin. PrCore also allows you to provide any additional information that can be used during the training and prediction stages of the plugin (if the plugin supports recognizing this information), which is particularly useful for custom-developed plugins, as non-default plugins may require more information to support their work. For more details, please refer to Provide Additional Information. Meanwhile, if you need to disable some plugins after creating a project so that they do not provide any suggestions and predictions when you provide new data, please refer to Disable or Enable Plugins.

Once the project is completed and trained, if you need to modify any column definitions or result definitions, you can redefine them using the method introduced in the Redefine Definitions section. At this point, the project will be retrained based on the modified information, and you will not need to upload any transaction logs again. Instead, if you want to keep all the parameter settings and additional information provided but upload a new log file, you can use the API in the Re-upload File section. In addition, if you want to delete the project, you can find the corresponding API in the Delete Project section. Want to know all existing projects? You can use the Get Projects API to get a list of all projects.

Since the project is fully trained, you can now submit new data to PrCore for prescriptions on ongoing cases. You have two options: the first is to follow the instructions in the Upload New Dataset section to send new events dataset to PrCore, where the dataset can only include ongoing cases, and then use the API in the New Dataset Results section to obtain the results for that dataset. The second option is to send new event data to the Input Data Stream API when new event data is available, so you can integrate your business management system to continuously send data to this API and obtain SSE-based result data streams using the method described in the Streaming Data Results section. If you have not yet integrated your business management system or written an automated data sending test script, you can use the method in the Simulate Data Stream section to have PrCore automatically simulate the reception of streaming data for you. At this point, you can directly obtain the results from the Streaming Data Results API to test the feature without additional programming.

By now, you have completed the main operations in the workflow and obtained the corresponding results.

If you are interested in how PrCore processes and annotates the datasets you provide based on the information you give, you can obtain the preprocessed data from PrCore through the API introduced in the Download Datasets section. Additionally, you can download the original files, test files containing only ongoing cases, and datasets used when PrCore simulates streaming data. These can help you verify whether the annotation and processing meet your expectations and provide more information for the development of your custom plugins.

One of the features of PrCore is its support for plugin integration: PrCore offers a simple and user-friendly plugin system, allowing you to create plugins based on specific algorithms that provide suggestions or predictions. You can explore the Development section for more information on core applications and plugin development details.