X hits on this document

PDF document

The split-fovea model of visual word - page 50 / 78

195 views

0 shares

0 downloads

0 comments

50 / 78

Chapter 2. The Simulations

45

a. Consistent Nonwords

Fixation Position Control Network Fixation Network

Change

1

4

6

+2

2

6

4

-2

3

4

4

+0

4

3

4

+1

5

6

7

+1

1

4

7

+3

2

5

8

+3

3

2

7

+4

4

10

4

-6

5

5

6

+1

Table 2.15: Nonword errors by fixation position for both consistent and inconsistent

nonwords. The last column indicates the difference between the control and the fixation

network.

Change

b. Inconsistent Nonwords

Fixation Position Control Network Fixation Network

Results

The overall results of the nonword testing can be seen in table 2.14.

The

upper part of the table shows the errors made on the consistent nonwords for both net- works and the lower part shows the errors made on the inconsistent nonwords. Both networks were able to pronounce most of the presented nonwords, both consistent and inconsistent, with the control scoring 85.2% on consistent and 85.1% on the inconsis- tent nonwords. The fixation net scored 83.9% and 81.7% respectively on the consistent and inconsistent nonwords. Again both networks are quite evenly matched with the fix- ation network getting only slightly more words wrong on both accounts (2 and 6 errors

more on the consistent and inconsistent nonwords respectively).

Document info
Document views195
Page views203
Page last viewedMon Dec 05 10:39:32 UTC 2016
Pages78
Paragraphs3087
Words17414

Comments