Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

Copy of The Medical Staff

No description

noor thamer

on 15 February 2015

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Copy of The Medical Staff

The Aim of Thesis
Bayesian Network
Thank You For All ِِ ِِِAttendees
Artificial Bee Colony

A pair (G;Q) where
G is DAG over variables Z, called network structure
Q is set of CPTs , one for each variable in Z, called network parametrization.
The challenge in this thesis lies in obtaining data that is acquired from the patient or the patient's file in order to take the medical history, medical tests of the patient, the symptoms suffered by the patient and entering these information into the system for the purpose of identifying the correct diagnosis and the accurate result for the patient as quickly as possible and with less cost.
(ABC) is an optimization algorithm based on the intelligent foraging behavior of honey bee swarm, proposed by Karaboga in 2005.
A hybrid approach based on Learning Bayesian Network Structures, is used were the best chosen structure through the hybrid algorithm is found. The hybrid MFBC-ABC model combines two algorithms: the Modified Full Bayesian Classifier algorithm (M-FBC) and Artificial Bees Colony algorithm (ABC). M-FBC is generated of ideal structure for each disease for access to be diagnosed, but we want to get into best structure by using ABC algorithm to obtain a more accurate diagnosis for an optimal structure. ABC has been selected because it is developed and based on nature-inspired ideas and it introduces several optimization algorithms. Most of swarm intelligent algorithms are meta-heuristic based search techniques and generally referred to as multipurpose optimization algorithms. The system (MFBC-ABC) is tested by the doctors in Iraqi hospitals. The accuracy for heart diseases is approximately (100%) and for nervous system diseases it is approximately (99%).
Conclusion BN
Advantages & Limitations BN
Code of Ethics for Physicians - "Medical Ethics"
Code of Ethics for Nurses
Created and revised by the AMA - American Medical Association,
pertains specifically to doctors.
Autonomy - respect of person, confidentiality
Beneficence - benefiting others
Non-Maleficence - "do no harm." respecting others
Justice, Truth, Honesty
Created and revised by the ANA -
American Nursing Association
Used in nursing education, emphasizes patient-centered care
Code of Ethics for Pharmacists
Examples of 2013 Ethical Issues
in the Medical Field
Birth Control
Gene Testing
Psychology Issues
Stem Cell Research
Created and revised by the APhA -
American Pharmacists Association
All ethical codes use principles from the Hippocratic Oath
Establishes the Pharmacist-Patient relationship
Established by worldview, social norms, truth
All patient data entered into the database to reduce files that written by hand.
The Aim (1)
The system is not needed all information about medical examinations to obtain the correct diagnosis.
It needs the patient's symptoms and medical history that mentioning when complained by patients.
The patient has not needed to visit more than one doctor to know the disease that he suffered.
Humans have been practicing medicine in one way or another for over a million years.
Advantages & Disadvantages
Bayesian reasoning
Gender ratio
ABC Algorithm
Step 1: Assign control parameters
Step 2: Initialize solutions
Step 3: Repeat until stopping criteria is met
Send the employed bee and calculate fitness
Send the onlookers and calculate fitness
Send the scout bees
Memorize the best solution
Step 4: Stop
Few control parameters
Fast convergence
Both exploration & exploitation
Search space limited by initial solution (normal distribution sample should use in initialize step)
Medical Estimation System Based on
Bayesian Decision Networks

Supervisor By

Asst. Prof. Dr. Ahmed Tarik
Prepare By

Noor Thamer Mahmood

The Aims of Thesis
The Challenge in This Thesis
Bayesian Networks
Artificial Bee Colony
The Block Diagram of Medical Estimation System (MES).
In this work a Bayesian approach for estimation system will be present to apply on a specific medical data. Also this proposed will be improve the Bayesian inference using an intelligent tool.

The proposed system is based on hybrid approach to medical diagnosis, which can help to diagnosis the diseases without needing all information about patient's medical history and symptoms, just asking about what the patients are feeling and as a result get diagnosed about any disease.

Video to Explain "How diagnose the disease"
The Aim (2)
The Aim (3)
The Cost
That leads to less cost time, effort and money.
The aims of this thesis is to present M-FBC model and MFBC-ABC model to obtain high speed of prediction for disease infected by the patients in order to save his life. This system services Iraqi society.
The Challenges
* Define network variables & their domain
Query & evidence variables, usually from problem statement
Intermediary variables, harder to determine
* Define network structure (edges)
What variables are direct causes of X?
* Define network CPTs
Determined objectively from problem statement
Reflection of subjective beliefs
Estimated from data
Constructing Bayesian Networks
* Probability theory well-established and well-understood.
* In theory, can perform arbitrary inference among the variables given a joint
* probability. This is because the joint probability contains information of all aspects of the relationships among the variables.

* Typically require initial knowledge of many probabilities…quality and extent of prior knowledge play an important role.
* Significant computational cost(NP hard task).
* Unanticipated probability of an event is not taken care of
There are four main types of reasoning in Bayesian networks:
• Causal reasoning (from H to E)
• Diagnostic reasoning / evidential reasoning (from E to H)
• inter-causal reasoning
• mixed type reasoning
Example BN
The Block Diagram of Medical Estimation System (MES)
DAG High Blood Pressure Disease
DAG Head Injury Disease

The MFBC-ABC is better than M-FBC, because that is the best for each disease as is generates a number of structures, by changing the values of the weights that have been calculated from the model (M-FBC). Re-arranged so that they appear different results. Choosing the best results are achieved on the best structure that is tested to get a better diagnosis of the disease. When tested (M-FBC) the accuracy for heart diseases is approximately (93%) and for nervous system diseases it is approximately (98%).
The success rate (accuracy) of system is (99.5) % knowing that if we want to reach better accuracy we must check data that have been taken from hospitals. This is done through the doctor. Who writes down all the data related to the patient (medical history, symptoms) in the patient's file.
Table (i), (i =1… N), where, N: no. of table.

A full BNs B, Probability (Prob).
Input All Tables in DB to M-FBC Model
Checked the Table in DB is ascertained to be empty.
Mutual information is calculate between each pair of attributes .
Threshold is calculate between each pair of attributes .
Find the maximum of threshold .
Weight is calculate for each attributes in the Table .
Sorted (descending) all weight for all variables.
The maximum weight is parent and the other are evidences.
CPT is calculate for each node in structure .
Probability is Calculate for Table .
Algorithm (M-FBC ) Structure.
M_FBC Structure
CPT for each node
Cont. Algorithm (MFBC - ABC) .

//Calculate Fitness (j) //
For each row in Test-Table (i),
Tested for B(i) new by calculate Probability .
If Accum_Prob >Accept_Accu then True= True +1
Else False = False +1
End for
Calculate Accuracy and put the result in Fitness [j].
End for
// Memorizing the best food source //

Selected the Maximum Accuracy from Fitness (j) and determined the position of the Maximum Accuracy.
Select the no. of Column (the position of the Maximum Accuracy) from FSM.
Create B(i) new from column (the position of the Maximum Accuracy) from FSM.
CPTs is calculated for each node in B(i) new by equation:
Save B(i) new.

End For
Algorithm (MFBC - ABC) .
Table (i), Test-Table (i), A Full BNs B(i), where, (i =1,…, M), M: no. of table, N: no. of columns (no. of Bees), Accept-Accu.

(Initialize the ABC and problem parameters)
A full BNs B(i) new, Accuracy (i). .
Input All Tables in DB to M-FBC Model

//Initialize FSM [SN, N], Fitness [N], where, SN = no. of node from full BNs//

B(i)old , N = no. of columns
- Make descending order all Weight (W).
- FSM[SN,0] = Weight Nodes for B(i) old
- Calculate Fitness
For each row in Test-Table (i) do
Tested for B(i) old by calculate Probability.
If Accum_Prob>Accept_Accu then True= True +1
Else False = False +1
End for
- Calculate Accuracy and put the result in Fitness[j].

//Employed Bee Phase//
For j = 1 to N
- For each row k in the columns make generated Wnew

for each neighboring weight .
Note that r ~ (0, 1) generates a uniform random number between 0 and 1.
End for
- Sorted (descending order) for each j in FSM.
- Create B(i) new for each j in FSM.
- CPTs is calculated for each node in B(i) new
The data are collected from some Iraqi hospitals. The data were acquired from the Statistics Division at Al-Shaheed Ghazi Hariri Hospital for Specialized Surgery and Al-Kindi Teaching Hospital. Patients' files are taken and information is extracted from it.
All data are taken from the patient's file, which was read by doctors, who helped me and taught me how to read the patient's file.
The data consists of medical history, symptoms, laboratory analysis … etc... The diseases that have been entered into the system are:
Data Acquisition
A- Heart Diseases

1. High Blood Pressure (I10).
2. Angina pectoris (I20).
3. Myocardial Infarction (I21).
4. Atrial Fibrillation (I48).
5. Heart Failure (I50).
B- Nervous System Diseases

1. Epilepsy (G40).
2. Nerve root and plexus compressions in diseases classified (G55)
3. Hydrocephalus (G91).
4. Head injury (G94).
5. Other diseases of spinal cord (G95).

1. eliminate the redundancy of data by identifying the patient's records that have the same attributes and unified it's in one record. Thus implied to take the 20 cases satisfactory and ratios repeat.

2. delete all the attributes that presents secondary symptoms and medical history. These deleted attributes are not useful because if they exist in the system guide it produces wrong diagnosis.
Filter and Features Extraction for Data
Table : Heart Diseases
Table : Nervous System Diseases
DAG High Blood Pressure Disease
DAG High Injury Disease
Full transcript