17
Semantic Medicine sedic

Sedic

Embed Size (px)

Citation preview

Page 1: Sedic

Semantic Medicinesedic

Page 2: Sedic

Project Objectives

Provide assistance in choosing a proper tea based on medical history and

a presumptive diagnosis made by system upon a set of symptoms

Provide assistance in following a cure, possible as an adjuvant treatment for

a disease.

Automatic exclusion from recommendations of certain teas that contain

substances to which the patient is allergic or suffer from a disease in which

are not commended.

Possibility to create a personal collection of teas, eventually with some

notes attached

Page 3: Sedic

Project Objectives

Provide a SPARQL end-point for unlimited interaction with data

Examples of utilization and a documentation will be provided

Ex: SELECT ?plant

WHERE{{?plant <http://http://myschema.org/terms/cure>

<http://myschema.org/terms/symptoms/coldness> }

}

Result: Tilia Tea known as : linden or basswood

Green Tea

Satureja hortensis knons as: Summer savory

Page 4: Sedic

Project Objectives

Provide a REST service through which a user can perform various tasks .

This provides the opportunity to integrate this system in other applications

The REST service will be accessible via URL

Ex. http://www.semanticmedicine.org/plants get all plants

http://www.semanticmedicine.org/plants/Tilia get tilia details

http://www.semnaticmedicinee.org/cure/cold get all plants that cure coldness

http://www.semanticmedicine.org/disease/kidney_stone get adjuvant

treatment for kidney stone

Page 5: Sedic

Presumptive Diagnosis

First step is to ask user to input some symptoms or diseases that is suffering

Based on this input:

User can ask for tea to treat his symptoms

User can ask for an adjuvant treatment for a suffering diseases

Page 6: Sedic

User input example

Example of input:

User suffers from a heart disease

Is not allergic to any substance

He wants a tea for his insomnia

The result in this case will recommend:

Matricaria chamomilla tea, also known as : chamomile, wild chamomile,

or scented mayweed

Valeriana officinalis tea, also known as: garden valerian,

garden heliotrope, or all-heal

Page 7: Sedic

Reminder and cure schedule

A user has the possibility to set a reminder that will help him in a cure with tea.

Ex: If he follows a cure with cherry tails for a kidney disease, he can

set a reminder that reminds him to drink his cherry tails tea.

Page 8: Sedic

Use case

Note:

User can ask for recommendations even if is not registered

This use case doesn`t include SPARQL end-point or REST service utilization!

Page 9: Sedic

Data source

Data will be collected using Apache Any23

The main source will be represented by DBpedia

Other sources will be used to collect additional information

Some information will be added manually, if is not possible in

another way

Page 10: Sedic

Data storage

Data will be storage as triples

In this purpose will be using Jena TDB

The final source will be represented by a union of graphs containing

common data. This process of union is automatically made by Jena

Page 11: Sedic

Example of data

Page 12: Sedic

Example of data

Page 13: Sedic

SPARQL end-point

Example: SELECT ?plant

WHERE{{?plant

<http://myschema.org/terms/cure>

<http://myschema.org/terms/symptoms/coldness> }

UNION {?plant

<http://myschema.org/terms/cure>

<http://myschema.org/terms/symptoms/insomnia>}

} ORDER BY ?plant

Result: Tilia tea also known as : linden or basswood

Page 14: Sedic

SPARQL end-point

Example: SELECT ?plant

WHERE{ {?plant

<http://myschema.org/terms/cure> <http://myschema.org/terms/symptoms/insomnia>}

MINUS {?plant

<http://myschema.org/terms/forbidden><http://myschema.org/terms/disease/high_blood_pressure>}

}

Result:

Matricaria chamomilla tea, also known as : chamomile, wild chamomile or scented mayweed

Valeriana officinalis tea, also known as: garden valerian, garden heliotrope or all-heal

Note: Recommendations will include Tilia (linden tea) due to high blood pressure

Page 15: Sedic

Architecture overview

SPARQL Query builder & Parser

UI HTML5 (with RDFa) & JS & CSS3

REST service

TDB Jena Triple store

SPARQL endpointForm Input interpreter

Page 16: Sedic

Data navigation

A visual navigation on data graph will be available with which users can

navigate through the plant, symptoms or diseases that are related

Useful tools: Sgvizler library

Similar with:

From: http://arborjs.org/

Page 17: Sedic

Conclusion

At the end of the project:

A user will be able to get useful information based on specific needs

Will have the possibility to be assisted to follow a cure

Possibility to interrogate data via a SPARQL end-point

A REST service will be provided

A visual exploration of data will be provided