Uploaded by БекФорд Зак

IT PROJECT CASE 2

advertisement
IT PROJECT CASE
Predict Early
Preparation
of Pharmacy
Prescriptions
Karimbaeva Ainur
|
Zhaksybay Bek
|
Bakytzhan Kalkaman
1.Company Overview
2.Challenge statement
3.Additional Information
Agenda
4.Proposed Solution
5.Project scope
6.Project Initiation
7.References
01
Prophet
Introduction
Company profile
Prophet is global consulting company that helps
customers with marketing, innovation, design and etc.
Many small and large companies has been acquired by
Prophet, such as:
In 2009 - Play company in Virginia
In 2012 - several agencies in Atlanta, Material Group in
Chicago & brand Figtree in Australia
In 2015 - Altimeter Group in California
In 2021 - KEYLENS Management Consultants in Munich
Headquarter
San Francisco, USA
Industry
Management Consulting
CEO
Michael Dunn
Number of employees
500-1000 people
Strategic Plan
Mission
The mission is focused on a higherorder approach to growth.
Values
Teamwork
Together we elevate each other's thinking and talents.
Fully Involvment
We love what we do and openly share our passion, time and
enthusiasm.
Vision
The team combines creativity, insight,
and technology in order to help
customers to unlock growth
Strong Bonds
We collaborate side by side, especially when the right way
turns out to be more risky.
Objectives
The organization purposed to keep a dynamic workplace that:
Helps its employees build satisfying and challenging careers
Is helpful for creative reasoning and arrangements
Perceives and remunerates greatness and accomplishment
Audience
Prophet Consulting's main target audience are companies or organizations
that are looking for practical and profitable solutions to stimulate growth
02
Prophet
Challenge Statement
Challenge Statement
As Prophet Consulting, we have to offer a solution to the
problem of Canadian Medicare related to patient satisfaction.
Our goal is to implement a mechanism that will predict the
diagnosis and pharmacy prescriptions in advance, thereby
significantly reducing the time spent in the hospital.
Definition of Done:
The time a patient spends in the hospital will decrease
Our AI based mechanism will predict the diagnosis and prescription with high accuracy
03
Prophet
Additional Information
Related Case
Using Machine Learning to Predict Early Preparation of Pharmacy
Prescriptions at PSMMC
Challenge - Long waiting times for the patient. One of the primary goals of key performance
indicators (KPIs) in health care is patient satisfaction. The studies found that patients spent too
much time waiting for a prescription, so they wanted to reduce the waiting time, which would
increase patient satisfaction over time.
Solution - Algorithm based mechanism. They created a mechanism based on machine learning
algorithms that can prescribe in advance regarding on patients' symptoms.
Results
During implementation, the "Weka" software tool has been used.
There were 1048575 instances in total, including 11 attributes.
Algorithm
Accuracy(%)
RT
97,22
J48
97,05
MLP
96,73
AdaBoost
96,16
04
Prophet
Proposed Solution
Implementation Strategy
1)Analyze and sort the historical data
2)Eliminate all empty data
3)Create Dataset/Database based on historical data
4)Develop Machine Learning model
5)Deploy machine learning model into production
6)Use the API on the hospital website to send and receive results
Association
Rule Learning
Algorithm
One of the most common data mining
approaches is the associative method,
which includes evaluating a collection of
properties from a database for repeated
relationships.
Apriori
To build association rules, this
algorithm works with a wide range
of datasets. It calculates the
itemset efficiently using a breadthfirst search and a Hash Tree.
To help us diagnose and prescriptions
in high accuracy, we will use the
following 4 metrics:
1)Support
2)Confidence
3)Lift
4)Confidence
03
Fullname
Fabiano Law
Kort Fortman
Izaak McLeary
Shawn Turmell
Fred Houston
Age
23
19
48
35
52
Phone Num.
87085180323
87018331035
87087831930
87069323103
87787439138
Gender
Symptoms
Diagnosis
Prescriptions
Frmale
Fever
Cough
Shortness of
breath
Coronavirus
Mukaltin 3 times a day;
Aspirin on an empty
stomach
Male
Shortness of
breath
Trouble
sleeping
Snoring
Obesity
Methmorphine before
bedtime
Male
Increased
hunger
Increased thirst
Weight loss
Diabetes
Methmorphine 2 times a day
Male
Fever
Cough
Fatigue
Coronavirus
Mukaltin 3 times a day
Coronavirus
Mukaltin 3 times a day;
Aspirin on an empty
stomach
Female
Fever
Cough
Loss of smell
1st Step: Support
Explanation:
This is a frequency based calculation
Formula:
Support(Coronavirus) = 3/5 = 0.6
Support(Obesity) = 1/5 = 0.2
Support(Diabetes) = 1/5 = 0.2
It means that the probability that the
patient is sick with coronavirus is
60%, obesity 20% and diabetes is
also 20% according to our data.
2nd Step: Confidence
Explanation:
This is a dependency based calculation
Formula:
Confidence(Coronavirus—>Mukaltin & Aspirin) = P(Coronavirus n Mukaltin & Aspirin) / Support(Coronavirus) = 2/5 : 3/5 = 2/3 ≈ 0.6
Confidence(Obesity—>Methomorphine) = P(Obesity n Methomorphine) / Support(Obesity) = 1/5 : 1/5 = 1
This means that the probability that a
patient will be prescribed both
mucaltin and aspirin for coronavirus
is 60%, and for obesity
methamorphine is prescribed in
100% of cases.
3rd Step: Lift
Explanation:
This is a dependency based calculation
Formula:
Lift(Coronavirus—>Mukaltin & Aspirin) = Confidence(Coronavirus—>Mukaltin & Aspirin) / Support(Mukaltin & Aspirin) = 2/3 / 2/5 = 5/3 ≈ 1.6
If lift > 1, mukaltin & aspirin are more likely to be prescripted together.
If lift = 1, there is no assiciation between mukaltin & aspirin.
If lift < 1, mukaltin & aspirin are not re likely to be prescripted together.
Programming
Language
Library
Framework
Python
Scikit-learn
Tensor Flow
13
The Jupyter Notebook is a free and opensource web platform that enables
developers to create and share
documents that include narrative text,
live code, graphics, and equations.
Machine learning (ML), numerical
simulation, data cleansing and
transformation, data visualisation, and
statistical modelling are all used in the
platform.
Algorithmia is an MLOps tool with which we
can use the AI created by our team on the
hospital website. To do this, run the ML
models and then host them on Algorithmia.
Then you can call our code as an API.
Methodology
XP emphasizes the technical elements of
software development, so we have chosen
it as our Agile approach. Following
engineering practises helps teams to
consistently provide high-quality code.
Pair Programming
Continuous Integration
Daily Deployment
Stories
03
05
Prophet
Project Scope
Benifits
Minimize the staff's
additional daily
workload
Reduce human
mistake to the
minimum
Using power of AI,
we can improve
processes and
customer
satisfaction
High-level
requirements
Structured data
Appropriate algorithm
Work platform
Deliverables
Machine Learning model
A system for diagnosing
New feature on the website
Risks
Risk
In the future,
Algorithmia may
change its policy
and make its
services fee-based.
Likelihood
Medium
Impact
Medium
Severity
Medium
Responsible
person
Mitigation Plan
Project Manager
We need to find other companies
that provide services for
deployment AI into production.
Alternatively, we can do it
ourselves.
Success factors
The algorithm correctly
diagnoses diseases
based on symptoms
with at least 80%
accuracy
Reduced time required
to get a prescription
06
Prophet
Project initiation
DATE
EVENT
PRIORITY
21.03.2022 20.04.2022
Building a team, reviewing available alternatives, and
selecting the best solution
High
20.04.2022 25.05.2022
Work on Association Rule Learning Algorithm
High
25.05.2022 10.06.2022
Test the Machine Learning model.
High
10.06.2022 21.07.2022
The solution has been fully implemented in the project.
High
Resource
plan
To resolve the problem, we will require fundamental resources
such as a workspace, office supplies, computers, and other staff
needs. In addition, we require IT expertise, a project manager,
and access to all Prophet data. According to preliminary
estimates, it will take 4 months.
Communications Plan
Stakeholder
Information
Delivery Mode
Frequency
Project manager
Report on
complate tasks
and work
process
Offline - Online
meetings
Once a week
Budget plan
Salary
№
Positions
Number of positions
Salary
Taxes
Total amount
1
Project manager
1
950
6
1007
2
Data scientists
3
1100
6
3498
3
ML engineers
2
1500
6
3180
4
F/B Dev
2
1000
6
2120
5
DevOps
1
570
6
604,2
6
Quality assurance engineers
1
700
6
742
Total (excluding tax)
9820
Total (including tax)
10409,2
Hardware
№
Producs
Number of pord
Price
Taxes
Toyal amount
1
Computer
23
450
12
11592
Total (excluding tax)
10350
Total (including tax)
11592
Furniture
№
Producs
Number of products
Price($)
Taxes($)
Toyal amount($)
1
Table
25
62
12
1736
2
Desk
10
70
12
784
3
Chair
25
100
12
2800
Total (excluding tax)
4750
Total (including tax)
5320
Stationery Needs
№
Products
Number of produc
Price($)
Taxes($)
Total amount($)
1
Stapler
5
0,52
12
2,91
2
A4-Paper
20
2,57
12
57,57
3
Marker
15
0,35
12
5,88
4
Notepad
35
0,67
12
26,26
5
Pen
65
0,17
12
12,38
6
Pencil
65
0,09
12
6,55
Total (excluding tax)
99,6
Total (including tax)
111,55
Quality Assurance plan
Quality plan
Experienced IT specialists will provide the quality. The project manager will guide them and keep an
eye on the quality of their work. Additionally, all pleasant circumstances will be established for staff,
as well as powerful devices, so that they may perform their duties as efficiently as possible.
Quality assurance activities
Meetings will be held to check the effectiveness of the task, and personnel will submit a report on
the work completed. Following the implementation of the solution, there will be a test period to
discover and eliminate bugs.
Quality standards
ISO 9001:2015
ISO/IEC 27001
Standard provides the standards for a quality
management system when a company:
Information security management
a) It must show that it can consistently deliver products and
services that fulfil consumer and statutory and regulatory
criteria, and
Though there are more than a dozen standards in the
ISO/IEC 27000 family, ISO/IEC 27001 is the most wellknown, establishing specifications for an information security
management system (ISMS). They allow any firm to manage
the security of assets such as financial data, intellectual
property, employee information, and information provided
by third parties.
b) works to promote customer satisfaction through effective
system application, including system improvement
procedures and verification of system adherence to customer
and relevant legislative and regulatory requirements.
07
Quebec & Ontario
Partnership between
2 provinces
Ontario needs partnership
"Beginning in 2017–18, the federal government reduced the growth of the
Canada Health Transfer. This decision removed $1.1 billion in federal support
to provinces and territories in 2017–18 alone".
(Provincial Govenment of Ontario)
"The provision of high-quality health care services to the people of Ontario is a
top priority for the Province, and should be for the federal government as well.
Compared to the 2017 Budget, Ontario is increasing investments in health care
by more than $5 billion between 2018–19 and 2020–21".
(Provincial Govenment of Ontario)
Implementation Strategy
1)Project initiation.
2)Preparation of the proposal.
3)Making a decision on the part of Ontario.
4)Decision to proceed with the project by both provincial governments.
5)Conclusion of an agreement.
6)Project implementation.
6.a)Ontario historical data will be given to Quebec to analyze and remove incomplete
data.
6.a)Use this data as the basis of an algorithm based mechanism.
7)Closure of the project.
References
Association Rule Learning - Javatpoint. (2018, May 21). Www.Javatpoint.Com.
https://shorturl.at/czDX1
CHI Software. (2022, January 7). Develop an Artificial Intelligence Software: How-to
Guide | CHI Software. Medium. https://shorturl.at/exCSY
Dobilas, S. (2022, February 5). Apriori Algorithm for Association Rule Learning —
How To Find Clear Links Between Transactions. Medium. https://shorturl.at/hotzT
Oppermann, A. (2022, February 9). Artificial Intelligence vs. Machine Learning vs.
Deep Learning: What’s the Difference? Built In. https://shorturl.at/lBDEM
Thank you
for
attention!
Have a great day ahead.
Download