prioritization-discovery 0.4.4

Python algorithm to discover, from an event log, prioritization rules that lead to one activity instance to be executed before another.

Prioritization Discovery

Python algorithm to discover, from an event log, the case priority levels and the rules that classify a process case in its corresponding level. For example, the cases of a process can belong to three priority levels (low, medium, high), where the activity instances of cases with high priority are executed before than activity instances of cases with low priority (when both of them are enabled at the same time).

Example of use

import pandas as pd

from prioritization_discovery.config import DEFAULT_CSV_IDS
from prioritization_discovery.discovery import discover_priority_rules

# Read event log
event_log = pd.read_csv("path_to_event_log.csv")
event_log[DEFAULT_CSV_IDS.enabled_time] = pd.to_datetime(event_log[DEFAULT_CSV_IDS.enabled_time], utc=True)
event_log[DEFAULT_CSV_IDS.start_time] = pd.to_datetime(event_log[DEFAULT_CSV_IDS.start_time], utc=True)
event_log[DEFAULT_CSV_IDS.end_time] = pd.to_datetime(event_log[DEFAULT_CSV_IDS.end_time], utc=True)
# Get priority levels and their rules
case_attributes = discover_priority_rules(
    event_log=event_log,
    attributes=['loan_amount', 'client_type']  # Case attributes to consider in the rule discovery
)

To see a more detailed example of use, and the format of the output, you can check this test file.

No enabled time available

To identify which activity instances have been prioritized over others, the information of the enabled time has to be available in the event log. In case it is not available, consider using this Python library to estimate them.