Team:Prairie View/Modeling

From 2008.igem.org

(Difference between revisions)
Line 15: Line 15:
<br>
<br>
<br>
<br>
-
 
+
|[[Image:Chart1.jpg|600px|center|]]<br>
<h3>Training using back propagation</h3>:<br>
<h3>Training using back propagation</h3>:<br>

Revision as of 21:45, 29 October 2008


Home The Team The Project Parts Submitted to the Registry Modeling Notebook


Contents

Artifical Neural Network Modeling of the Molecular Biosensor



|
Chart1.jpg

Training using back propagation

:

Select function in matlab library to fit data

Find error, and compare to target error

General error function


Finally select function that gives least error

Sigmoid function

An enose is an analytic device originally used for detecting chemicals and their concentrations in vapors


Using E-Nose allow us to apply it to a metal sensor by finding the functional relationship, which is the response to the concentration and type of metal




Now that we have built our Molecular Sensor , with E-Nose we can:
Use experimental data for individual metal ion protein sequence, and ligations

Look at a wider range of variation in the concentrations

Collect data from rejected samples to determine the reliability of network

And use the completion of the final network to identify the ion as well as it’s corresponding concentration













Home The Team The Project Parts Submitted to the Registry Modeling Notebook