lections № 1

55
Lections Lections 1 1 Decision Support and Expert Systems in Medicine

Upload: rashad-stokes

Post on 01-Jan-2016

72 views

Category:

Documents


1 download

DESCRIPTION

Lections № 1. Decision Support and Expert Systems in Medicine. Main Questions. Decision Support Systems Basics . Decision Support Systems in Medicine Expert systems Artificial neural network. 1. Decision Support Systems Basics. Information systems definition - PowerPoint PPT Presentation

TRANSCRIPT

LectionsLections № №11

Decision Support and Expert Systems in Medicine

Main QuestionsMain Questions

Decision Support Systems Basics.Decision Support Systems in

MedicineExpert systemsArtificial neural network

1.1. Decision Decision SupportSupport Systems Systems BasicsBasics

Information systems definitionDecision support systems definitionDSS TaxonomiesDSS Architecture Benefits of DSS

1.1.11. . Information system definitionInformation system definition

An Information System (IS)Information System (IS) is the system of persons, data records and activities that process the data and information in a given organization, including manual processes or automated processes;

The computer-based information systems are the field of study for Information Information technologies (IT) technologies (IT)

1.1.22.. Decision Decision supportsupport systems systems definitiondefinitionDecision support systems (DSS)Decision support systems (DSS) are a

class of computer-based information information systemssystems including knowledge based knowledge based systemssystems that support decision making decision making activitiesactivities..

The term decision support systemdecision support system has been used in many different ways and has been defined in various ways depending upon the author's point of view:

1.1.22.. Decision Decision supportsupport systems systems definitiondefinition DSSDSS is a model-based set of procedures for

processing data and judgments to assist a manager in his decision-making..– Little, J.D.C.(1970, April). "Models and Managers:The Concept of a

Decision Calculus.

DSS DSS is an extendible systems capable of supporting ad hoc data analysis and decision modeling, oriented toward future planning, and used at irregular, unplanned intervals. – Moore, J.H.,and M.G.Chang.(1980,Fall)."Design of Decision Support

Systems."

1.1.22.. Decision Decision supportsupport systems systems definitiondefinition DSS DSS it is a computer-based system that aids the

process of decision makingdecision making.. – Finlay, P. N. (1994). Introducing decision support systems.

DSS DSS it an interactive, flexible, and adaptable computer-based information systeminformation system, especially developed for supporting the solution of a non-structured managementmanagement problem for improved decision makingdecision making. It utilizes data, provides an easy-to-use interfaceinterface, and allows for the decision maker's own insights. – Turban, E. (1995). Decision support and expert systems: management

support systems.

1.1.33.. DSS DSS TaxonomiesTaxonomies

Using the relationship with the user as the criterion can be differentiate passive, active, passive, active, and cooperative DSScooperative DSS:– Haettenschwiler, P. (1999). Neues anwenderfreundliches Konzept der

Entscheidungsunterstützung.

A passive DSSpassive DSS is a system that aids the process of decision making, but that cannot bring out explicit decision suggestions or solutions.

An active DSSactive DSS can bring out such decision suggestions or solutions.

1.1.33.. DSS DSS TaxonomiesTaxonomies

A cooperative DSScooperative DSS allows the decision maker (or its advisor) to modify, complete, or refine the decision suggestions provided by the system, before sending them back to the system for validation. The system again improves, completes, and refines the suggestions of the decision maker and sends them back to her for validation. The whole process then starts again, until a consolidated solution is generated.

1.1.33.. DSS DSS TaxonomiesTaxonomies Using the mode of assistance as the criterion,

differentiates communication-driven DSS, communication-driven DSS, data-driven DSS, document-driven DSS, data-driven DSS, document-driven DSS, knowledge-driven DSS, and model-driven knowledge-driven DSS, and model-driven DSS:DSS:– Power, D. J. (2002). Decision support systems: concepts and

resources for managers.

A communication-driven DSScommunication-driven DSS supports more than one person working on a shared task; examples include integrated tools like Microsoft's NetMeeting

1.1.33.. DSS Taxonomies DSS Taxonomies A data-driven DSSdata-driven DSS or data-oriented DSS

emphasizes access to and manipulation of a time seriestime series of internal company data and, sometimes, external data.

A model-driven DSSmodel-driven DSS emphasizes access to and manipulation of a statistical, financial, optimization, or simulation model. Model-driven DSS use data and parameters provided by users to assist decision makers in analyzing a situation; they are not necessarily data intensive.

1.1.33.. DSS Taxonomies DSS Taxonomies A document-driven DSSdocument-driven DSS manages, retrieves

and manipulates unstructured information in a variety of electronic formats.

A knowledge-driven DSSknowledge-driven DSS provides specialized problem solvingproblem solving expertise stored as facts, rules, procedures, or in similar structures.– Moust important for medical applications.

1.1.33.. DSS Taxonomies DSS TaxonomiesThe DSSDSS has been classified into the following six

frameworks: Text-oriented DSSText-oriented DSS;; Database-oriented DSSDatabase-oriented DSS;; Spreadsheet-oriented DSSSpreadsheet-oriented DSS;; Solver-oriented DSSSolver-oriented DSS;; Rule-oriented DSSRule-oriented DSS;; Compound DSSCompound DSS (hybrid system ).

– Holsapple, C.W., and A. B. Whinston. (1996). Decision Support Systems: A Knowledge-Based Approach.

1.1.33.. DSS Taxonomies DSS Taxonomies

1.1.44.. DSS DSS ArchitectureArchitecture

The DSSThe DSS

Data Data Management Management ComponentComponent

Model Model Management Management ComponentComponent

User Interface User Interface Management Management ComponentComponent

1.1.44.. DSS DSS ArchitectureArchitecture The Data Management Component (DBMS)Data Management Component (DBMS)

stores information. Information Information can be further subdivided into:– derived from an organization's traditional data

repositories,

– derived from external sources such as the Internet,

– or derived from the personal insights and experiences of individual users;

1.1.44.. DSS DSS ArchitectureArchitecture the Model Management ComponentModel Management Component handles

representations of events, facts, or situations (using various kinds of modelsmodels, two examples being optimization modelsoptimization models and goal-seeking goal-seeking modelsmodels);

the User Interface Management ComponentUser Interface Management Component is of course the component that allows a user to interact with the system.

1.1.55.. Benefits of DSSBenefits of DSS

1. Improving Personal Efficiency 2. Expediting Problem Solving 3. Facilitating Interpersonal

Communication 4. Promoting Learning or Training 5. Increasing Organizational Control

2. 2. Decision Decision SupportSupport Systems Systems in Medicinein Medicine

Digital dashboardClinical decision support systemMedical logic module. Arden syntax

22..1. 1. Digital dashboardDigital dashboardA digital dashboarddigital dashboard (enterprise dashboard (enterprise dashboard

or executive dashboard) executive dashboard) is a business management tool used to visually ascertain the status (or "health") of a business enterprise via key business indicators.

Digital dashboards use visual, at-a-glance displays of data pulled from disparate business systems to provide warnings, action notices, next steps, and summaries of business conditions.

22..1.1. Digital dashboard Digital dashboardSome benefits to using digital dashboards

include:Visual presentation of performance measures Elimination of duplicate data entry. Ability to identify and correct negative trends. Measure efficiencies/inefficiencies. Ability to generate detailed reports showing

new trends. Increase overall revenues. Ability to make more informed decisions

based on collected BI (business intelligence) Align strategies and organizational goals.

22..1. 1. Digital dashboard screenshotDigital dashboard screenshot

22.2.2. . Clinical decision support systemClinical decision support systemClinicalClinical (or diagnosticdiagnostic) decision decision

support systems (CDSS)support systems (CDSS) are interactive computer programs, which are designed to assist physicians and other health professionals with decision making tasks.

"Clinical Decision Support systemsClinical Decision Support systems link health observations with health knowledge to influence health choices by clinicians for improved health care".– Dr. Robert Hayward of the Centre for Health Evidence

22.2.2. . Clinical decision support systemClinical decision support systemThe basic components of a CDSS include: a dynamicdynamic (medical) knowledge baseknowledge base an inferencing mechanisminferencing mechanism (usually a set

of rules derived from the experts experts and evidence-based medicineevidence-based medicine) and implemented through medical logic medical logic modulesmodules based on a language such as Arden syntaxArden syntax.

It could be based on Expert systemsExpert systems or artificial neural networksartificial neural networks or both (connectionist expert systems).

22.2.2. . Clinical decision support systemClinical decision support system

2.2.33.. Medical logic module Medical logic module A medical logic module (MLM)medical logic module (MLM) is an

independent unit in a health knowledge base that combines the knowledge required and the definition of the way it should be applied for a single health decision.

An event monitor program in an electronic electronic medical record (EMR)medical record (EMR) uses it, on occurrence of defined conditions.

A grammar - the Arden syntaxArden syntax has been defined which would make MLMs swappable between different hardware and software platforms.

2.3. MLM. Arden syntax 2.3. MLM. Arden syntax The Arden syntaxArden syntax is a grammar for

describing medical conditions and recommendations, used in Medical algorithms.

MLMMLM are written in Arden syntaxArden syntax, and are called by a program - an event monitor - when the condition they are written to help with occurs.

Arden syntax was formerly a standard under ASTMASTM, and is now part of HL7HL7.

2.3. MLM Example2.3. MLM Examplemaintenance: title: Creatinine clearance;; version: 1.09;; author: George Hripcsak, M.D.;;library: purpose: To calculate the creatinine clearance for every timed urine

collection;; explanation: When a timed urine collection is stored, the MLM checks for;;knowledge: data: let urine_creat_storage be event {'32506','1762'}; let (urine_creat, collect_time) be read last {'evoking',

'dam'="PDQRES1";'1762'; '1537'};;; evoke: starting time of urine_creat_storage;; logic: let serum_creat be nearest (time of urine_creat) from (serum_creat_list

where it is number); let creat_clear be 0.07 * (24 / collect_time) * (urine_creat / serum_creat); conclude true; ;; action: write "The creatinine clearance is " ||int(0.5+creat_clear)|| " ml/min

based upon a " ||collect_time|| " hour urine creatinine of " ||urine_creat||.....; ;;end:

33. . Expert systemsExpert systems

Expert systems definitionArchitecture of the ESES Advantages and disadvantages

3.1. Expert system definition3.1. Expert system definition

An expert systemexpert system, also known as a knowledge based systemknowledge based system, is a computer program that contains the knowledge and analytical skills of one or more human experts, related to a specific subject.

This class of program was first developed by researchers in artificial intelligenceartificial intelligence during the 1960s and 1970s and applied commercially throughout the 1980s.

3.1. Expert system definition3.1. Expert system definition

Expert systems provide expert-quality expert-quality adviceadvice, diagnoses and recommendations on real world problems

Designed to perform function of a human human expertexpert

Examples:

– Medical diagnosis - program takes place of a doctor; given a set of symptoms the system suggests a diagnosis and treatment

3.1. Prominent medical ES3.1. Prominent medical ES CADUCEUS (expert system)CADUCEUS (expert system) - Blood-borne

infectious bacteria. MycinMycin - Diagnose infectious blood diseases

and recommend antibiotics (by Stanford University)

STD WizardSTD Wizard - Expert system for recommending medical screening tests

Dendral Dendral - Analysis of mass spectra

3.2. Architecture of the ES3.2. Architecture of the ES

User InterfaceUser Interface

Knowledge Knowledge BaseBase

Inference Inference EngineEngine

Working Working MemoryMemory

Production Production rulesrules

Recognise-Recognise-act cycleact cycle

Compared to Compared to production production

rulesrules

3.2. ES - Introduction to Rules3.2. ES - Introduction to Rules The knowledge baseknowledge base of an expert system is often

rule basedrule based – the system has a list of rules which determine what should be done in different situations

These rules are initially designed by human experts

The rules are called production rulesproduction rules Each rule has two parts, the condition-action condition-action

pairpair1.1. Condition Condition – what must be true for the rule to

fire2.2. ActionAction – what happens when the condition is

met Can also be thought of as IF-THEN rulesCan also be thought of as IF-THEN rules

3.2. ES - Conditions example3.2. ES - Conditions example Conditions are made up of two parts:

– ObjectsObjects – eg the weather

– The objects’ valuevalue – eg sunny IF sunny(weather) THEN print “wear sunglasses”

May also be an operator, such as greater than:– IF >30(temperature) THEN print “take some water”

Conditions may also be joined together using AND, OR, NOT:AND, OR, NOT:– IF sunny(weather) AND outdoors(x) print “take your

sunglasses x”

3.2. ES - The Working Memory3.2. ES - The Working Memory

The contents of the working working memorymemory are constantly compared to the production rules

When the contents match the condition of a rule, that rule is fired, and its action is executed

More than one production rule may match the working memory

3.2. ES - The Working Memory3.2. ES - The Working Memory

The system cycles around in the recognise-act cyclerecognise-act cycle

Whenever a condition is matched, it is added to the conflict setconflict set – all the rules which are currently matched

The system must then decide which rule within the conflict set to fire – conflict conflict resolutionresolution

3.2. ES - Knowledge acquisition3.2. ES - Knowledge acquisition The system must liaise with people in order to gain

knowledge and the people must be specialised in the appropriate area of activity.

The knowledge engineer (KE)knowledge engineer (KE) acts as an intermediary between the specialist and the expert system. The KE is also responsible for the self consistency of the data loaded.

Typical of the information that must be gleaned is vocabulary or jargon, general concepts and facts, problems that commonly arise, the solutions to the problems that occur and skills for solving particular problems.

3.2. ES - User Example3.2. ES - User ExampleThe end-user usually sees an expert system through an

interactive dialoginteractive dialog, an example of which follows:– System:System: What is the patient's name?– User:User: George Washington– System:System: What is the patient's age?– User:User: 252 years– System:System: Is the patient male or female?– User:User: Male– System:System: What are the symptoms?– User:User: Headache dizziness upset stomach high fever– System:System: When did these symptoms first appear?– User:User: December 13,1799– System:System: I conclude that the patient has a viral infection , my

recommendation is that he should take two aspirin, drink plenty of fluid, get lots of rest and call me in the morning.

In expert systems, dialogs are not pre-planned. There is no fixed control structurecontrol structure. Dialogs are synthesized from the current information and the contents of the knowledge base

3.3. ES Advantages and disadvantages3.3. ES Advantages and disadvantages AdvantagesAdvantages: Provides consistent answers for repetitive decisions,

processes and tasks Holds and maintains significant levels of information Encourages organizations to clarify the logic of their

decision-making Never "forgets" to ask a question, as a human mightDisadvantagesDisadvantages: Lacks common sense needed in some decision making Cannot make creative responses as human expert would

in unusual circumstances Domain experts not always able to explain their logic and

reasoning Errors may occur in the knowledge base, and lead to

wrong decisions Cannot adapt to changing environments, unless knowledge

base is changed

44. . Artificial neural networkArtificial neural network

Artificial neural network definitionNetwork model in artificial neural networkLearning of the ANNAn ANN Application

4.1. Artificial neural network4.1. Artificial neural network definitiondefinition

An artificial neural network (ANN)artificial neural network (ANN) is a mathematical modelmathematical model or computational modelcomputational model based on biological neural networksbiological neural networks.

It consists of an interconnected group of artificial artificial neuronsneurons and processes information using a connectionist connectionist approach to computationcomputation.

In most cases an ANN is an adaptive systemadaptive system that changes its structure based on external or internal information that flows through the network during the learning phase.

4.1. ANN4.1. ANN exampleexample

A neural network is an interconnected group of nodes, akin to the vast network of neurons in the human brain.

4.1. ANN4.1. ANN exampleexample

Component based representation of a neural network. This kind of more general representation is used by some neural network neural network

softwaresoftware

4.1. ANN4.1. ANN exampleexample

Currently, the term Artificial Neural Network (ANN)Artificial Neural Network (ANN) tends to refer mostly to neural network models employed in statistics, cognitive psychologystatistics, cognitive psychology and artificial intelligenceartificial intelligence.

Neural networkNeural network models designed with emulation of the central nervous system (CNS)central nervous system (CNS) in mind are a subject of theoretical neuroscience (computational theoretical neuroscience (computational neuroscience).neuroscience).

In modern software implementationssoftware implementations of artificial neural networks the approach inspired by biology has more or less been abandoned for a more practical approach based on statistics statistics and signal processingsignal processing.

4.2. Network model in ANN4.2. Network model in ANNANN models are essentially simple

mathematical modelsmathematical models defining a function:

The function f(x)f(x) is defined as a composition of other functions ggii(x)(x),, which can further be defined as a composition of other functions.

This can be conveniently represented as a network structurenetwork structure, with arrows depicting the dependencies between variables.

4.2. Network model in ANN4.2. Network model in ANNA widely used type of composition is the

nonlinear weighted sumnonlinear weighted sum:

where KK is some predefined function, such as the hyperbolic tangent or other.

It will be convenient for the following to refer to a collection of functions ggii as simply a vector:

4.2. Network model in ANN4.2. Network model in ANN

Figure depicts such a decompositiondecomposition of ff, with dependencies between variables indicated by arrows. These can be interpreted in two ways:

The functional viewfunctional view: the input xx is transformed into a 3-dimensional vector hh, which is then transformed into a 2-dimensional vector gg, which is finally transformed into ff. This view is most commonly encountered in the context of optimizationoptimization.

The second view is the probabilistic view: the random random variablevariable F = f(G)F = f(G) depends upon the random variable G = G = g(H)g(H),, which depends upon H = h(X)H = h(X),, which depends upon the random variable XX. This view is most commonly encountered in the context of graphical modelsgraphical models.

4.3. Learning4.3. Learning of the ANNof the ANNThe most interest in ANN is the possibility of learninglearning : Given a specific task to solve, and a class of functions

FF, learning means using a set of observationsobservations, in order to find which solves the task in an optimal optimal sensesense.

This entails defining a cost function:cost function: The cost function CC is an important concept in learning,

as it is a measure of how far away we are from an optimal solution to the problem that we want to solve.

Training a neural network model essentially means selecting one model from the set of allowed models that minimises the cost criterion.

4.4. An ANN Application4.4. An ANN ApplicationThe tasks to which artificial neural networks are

applied tend to fall within the following broad categories:

Function approximationFunction approximation, or regression regression analysisanalysis, including time series predictiontime series prediction and modelingmodeling.

ClassificationClassification, including patternpattern and sequence recognitionsequence recognition, novelty detectionnovelty detection and sequential decision making.

Data processingData processing, including filtering, clustering, blind source separation and compression.

4.4. An ANN Application4.4. An ANN ApplicationANN Application areas include: system identification and controlsystem identification and control (vehicle control,

process control); game-playing and decision makinggame-playing and decision making (backgammon,

chess, racing); pattern recognitionpattern recognition (radar systems, face identification,

object recognition and more); sequence recognitionsequence recognition (gesture, speech, handwritten

text recognition); medical diagnosis and visualizationmedical diagnosis and visualization; financial applicationsfinancial applications (automated trading systems); data miningdata mining (or knowledge discovery in databases,

"KDD").

ConclusionConclusion

In this lecture was described next questions:

Decision Support Systems Basics.Decision Support Systems in

MedicineExpert systemsArtificial neural network

LiteratureLiterature

Electronic documentation on to the TDMU server:http://www.tdmu.edu.te.uahttp://www.tdmu.edu.te.ua