SAMPLE_KDU-SC- AI &DL POSTER

1.

Real-Time Face Detection and Gender Prediction System Using Machine Learning with
Webcam Integration
1
Mohammed, Mohammed Abdulhakim Al-Absi
1,2
2
Department of Smart Computing, Kyungdong University, 46 4-gil, Bongpo, Gosung
, Gangwon-do 24764, Korea
Supervisor: Dr. Mohammed A.A
Development Tools
Game Overview
With the aim to demonstrate and raise the usefulness of process
mining analysis in the real word, we describe an application of
process mining in a real material purchasing process in a ship and
naval parts manufacturing company in Korea. This process is
characterized by a long lead time.
Under pressure of the fast growth of ICT, organizations are required
to continuously enhance their business processes to defend their
market position.
To achieve this,
Process Mining has emerged
as a mean to analyze the behavior
of companies by extracting process
related knowledge from events logs
stored in today’s Information
Systems
•The company adopts a Shipbuilding Pro
cessing Plan Management (SPPM) Syste
m to monitor its processes.
However, the company has no
clear idea of what is really
happening in the material
purchasing process and why Discover the
the lead time of this process is actual AS-IS
process
taking too much time.
Process mining overview
Process mining which falls between data mining, artificial
intelligence, process modeling, and analysis, forms a new form of
process-driven analytics.
Case Study
Result
Analyze
process
performance
Process
Flow Diagram
Method
-The framework of the Analysis consists of four steps: data
extraction, data pre-processing,
interpretation of result.
purchasing
process
analysis,
The final obtained event log after the
pre-processing contains 663 cases and
9829 complete events recorded within
one year and seven months.
We have mined the model of material
purchasing process using fussy miner
technique due to its strong ability in
handling less structured process
models.
Discovered model contains 9 activities
and the thickness of arcs and the
coloring depict how frequently each
activity and path have been executed.
Simulation Result And Analysis
Interpretation
Performance Analysis
Method
Here, the discovered model is extended
with the timestamps of both activities
and between activities. With the time
attribute of events, bottlenecks which
negatively affect the performance of the
whole process can be identified.
Result 1
Most of the activities are taking few
minutes, except the activities purchase
approval verification by purchase and
business
departments,
purchase
finalization and purchase orders which
are taking a long time compared to oth
er activities (38.1, 34.6, 22.8 and 15.1
hours respectively).
Result 2
Concerning the time between activities,
we have identified several backward
works that are taking too much time.
Based on the identified activities that are taking too much time,
the reworks, and the time spent before the reworks are
performed, we can say that too many delays are occurring in
the material purchasing process which negatively affect its
performance.
Since the production process strongly depends on the availabilit
y of materials, any delay occurring in the material purchasing
process engenders the delay of the production process.
Efficient actions need to be undertaken to reduce the
backward executions and minimize the time of the identifie
d activities that are taking too much time. The waiting time
to restart the process also needs to be reduced.