'Process and progress for natural language analysis of company communication?
Assume there is a large record of all different kinds of inter-employee and customer communications (e.g. mails, chat transcripts, OCRed letters) which should be analyzed, for sentiment analysis.
I am trying to explain my customers how NLP works using BPMN. The challenge is to find an appropriate level of detail. Decision makers tend to see text analytics as a magic wand which delivers reproducible, object answers immediately. I somehow want to emphasize the iterative nature of the whole process. BPMN seems a nice way of doing so, because one could visualize the process of our efforts: If I have a huge dataset, I could have tokens at different tasks, showing how many texts of the corpora are already at what stage, and this even can be automatically.
Before reinventing the wheel: Has anybody already married NLP and BPMN in some way? Are there any major flaws in my diagram? Is there a better way (maybe avoiding BPMN) to describe the process and to measure progress?
Source code for the diagram: https://pastebin.com/raw/zMzuLvJH (feel free to use or modify)
<?xml version="1.0" encoding="UTF-8"?>
<bpmn:definitions xmlns:bpmn="http://www.omg.org/spec/BPMN/20100524/MODEL" xmlns:bpmndi="http://www.omg.org/spec/BPMN/20100524/DI" xmlns:di="http://www.omg.org/spec/DD/20100524/DI" xmlns:dc="http://www.omg.org/spec/DD/20100524/DC" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" id="Definitions_1" targetNamespace="http://bpmn.io/schema/bpmn" exporter="Camunda Modeler" exporterVersion="1.8.2">
<bpmn:process id="Process_1" isExecutable="false">
...
References
- Set of rules for textual analysis - Natural language processing
- Julia Silge and David Robinson: Text Mining with R - A Tidy Approach
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|