A company wants to effectively manage its management system, making it a bigger role, a workflow management system is very necessary, through such a system, we can perform the data of the currently running workflow Effective monitoring and management.
However, many companies have had one or more daily management systems a long time ago. For example, the office personnel have their own OA system, the production department has its own production and management system, the sales department has a sales management system, etc., each system may have its own platform, their own database. If there is a workflow management system, it is possible to effectively, uniform management, so you can make better operation of each system. However, in the process of the most basic step - during workflow modeling, we will encounter great difficulties.
Basic methods We might want to analyze all business processes, then extract reasonably, and use it as the basis for establishing a model, and then you can start modeling.
There are many works to be handmade, and work efficiency is imagine that if we can use the tool automatically extract the necessary models from existing data, all the work is automated, and then manually The local adjustment is greatly improved.
However, the data inside is not in the database can be used directly, because there may be some data necessary to make a model mining, so we have to perform related processing.
1 . Pat out the definition of data flowing in the process
2. Find related information in the database
3. Eliminate data we have not used, add the necessary additional information
Among them, useless data should be some details of the details. Since we are going to do, the data of many special details is meaningless to us. The data that needs to be added should be a timestamp, successful failure information, the last link, the next link, etc., these data may be added while adding, or it may be added together, need to be adjusted according to different situations. .
Next, we have the source data that excavated, you can make a model mining.
From these rich data, we can dig out several types of data
1) Model for normal flow
2) Model of abnormal processes
3) Model for error processes
Then through the analysis of the first two data, we can get a workflow model.
