Difference between revisions of "Workflow : Information Extraction From Query"
Anote2Wiki (talk | contribs) |
Anote2Wiki (talk | contribs) |
||
Line 5: | Line 5: | ||
The Information Extraction From Query Workflow allows you to set up some tasks over the results of a query, including Corpus creation (mandatory) and NER and RE Processes (optionally). | The Information Extraction From Query Workflow allows you to set up some tasks over the results of a query, including Corpus creation (mandatory) and NER and RE Processes (optionally). | ||
− | |||
To run the workflow, you must select a Query object on the Clipboard and right click '''Workflow -> Information Extraction (From query)''' or click on the '''Workflow Information Extraction''' button in the Query View. | To run the workflow, you must select a Query object on the Clipboard and right click '''Workflow -> Information Extraction (From query)''' or click on the '''Workflow Information Extraction''' button in the Query View. |
Revision as of 13:24, 16 April 2013
Contents
Operation
The Information Extraction From Query Workflow allows you to set up some tasks over the results of a query, including Corpus creation (mandatory) and NER and RE Processes (optionally).
To run the workflow, you must select a Query object on the Clipboard and right click Workflow -> Information Extraction (From query) or click on the Workflow Information Extraction button in the Query View.
Select Query Publications
The first step of the workflow is to select candidate publications to create a Corpus. If you have already selected publications using the Query View this information will already be present. After selecting the desired publications, press the Next
Select Steps
The next step is to determine the steps that will be executed by the workflow. The Corpus creation is mandatory but NER and RE processes can also be applied to this Corpus in subsequent operations. After selecting the steps, press next to continue.
Create Corpus
(Mandatory)
The next step is to configure the corpus creation, where you have to select the name of the Corpus and its type.
Corpus Type
Abstract : : Only publications with abstracts will be considered.
Full Text : Only publications with full Text / PDF will be considered.
Retrieve PDF : Only publications with full Text / PDF will be considered, and a Journal Retrieval Process will be launched to collect all selected documents
If you just have selected corpus creation, the processing of your data will start after clicking ok.
When this process finishes, you can see the results in the Workflow_Report
Otherwise, if other steps are included, press next to proceed to NER configuration.
Select NER Process
(Optional)
The next step is the configuration of an NER Process. Using the combo box, select the NER process configuration that is more suitable.
NER - Based in Lexical Resources
You can select a NER Based in lexical resources. The configuration has two panels: Basics an Advanced:
Basic
Here you can select one or more dictionaries to use as resources in the NER.
Advanced Option
Expert users can configure some advanced options. These options are based in NER - Lexical Resource: check details in NER - Lexical Resources
If your workflow terminates in the NER process, the data processing will start after clicking ok.
When the process finishes the processing, you can see results the in the Workflow_Report
Otherwise, if an RE process will be conducted, press next to proceed to its configuration.
Select RE Process
(Optional)
The next step is the configuration of an REProcess. Using the combo box select the RE process that best serves your needs.
RE Based in POS-Tagging
You can select an RE Based in Natural Language processing. The configuration has two panels: Basic an Advanced.
Basic
Here you can select the relation model
Advanced
Expert users can configure some advanced options. These options are based in the RE operations detailed in: Relation Extraction
Processing
As the process proceeds you can check the status of activity looking at the progress bar
Workflow Report
As a result of the process, a Workflow_report is shown.