So you are trying to use a neural network to recognize noise events ...
You must follow these steps to get your job done.
- Prepare a text file as explained next in this page and save it on you local disk with .txt extension.
- Load the text file you prepared from the applet using the 'Load events' button and wait ...
You will get as output a floating point value between 0.0 and 1.0 that is the aeronautical probability of the event clamped manually to the neural network, more you will have an estimation of the lvaj value if you clamp to the network all the events contained in the file you choose.
The text file format
You need the following informations about each noise event:
-
The date and time of the noise event beginning instant:
- e.g.
2009-04-03 12:56:34
- e.g.
-
The duration in seconds of the noise event:
- e.g.
21
- e.g.
-
The LMAX value of the noise event:
- e.g.
65.6
- e.g.
-
The LAEQ of the noise event:
- e.g.
62.1
- e.g.
-
The SEL of the noise event:
- e.g.
75.4
- e.g.
-
The SEL 10 of the noise event:
- e.g.
75.4
- e.g.
Each one of these values (from 1 to 6) must be separated by a single space character:
e.g. 2009-04-03 12:56:34 21 65.6 62.1 75.4 75.4
and must be a single line in your input file, so you will end up with something like this:
...
2009-06-01 00:13:55 16 91.0 80.5 92.6 83.4
2009-06-01 06:33:52 41 76.0 70.6 86.8 84.2
2009-06-01 06:47:25 21 65.6 62.1 75.4 75.4
2009-06-01 06:56:15 38 70.0 66.3 82.1 81.7
2009-06-01 06:59:38 48 72.4 67.8 84.6 82.8
2009-06-01 07:00:59 15 82.1 78.7 90.4 90.1
...
Take this brief example file here.
