| Summary: | 碩士 === 臺中師範學院 === 數學教育學系 === 90 === According to the “psyehosocial development theory” advocated by Erikson(1963), the development target of adolescents is to conquer “idenity crisis” to reach “self-identify”. It is the key of a life evolution that whether arrivals “self-identify” or not during juvenile period.
Recently, neural networks have attracted by various domain experts. Further more; the population of multiprocessor systems takes advantages of the neural network, for example multi-processing and parallel–computing.
Basis on the degree of juvenile self- organization, it is important to teach adolescents, guide them to adapt society and achieve the self-integration. The goal of this research is how to integrate technologies of the neural network and developments of self-organization surveys by the prior fault-tolerance of neural network and abilities of self-learning.
By the research result, we can have following conclusion,
1. It is available way to apply the technology back-propagation neuron network to develop self-identity survey. After learning, the average error between the test sample and target is 4.63, and the result satisfied.
2. By the test of “Self-Identity Battery Ⅱ”，the average value of the degree of student’s identity is 53.9, the lowest is 33.348, the highest is 68.71 and the standard deviation is 7.30.
3. By the gender, the degree of student’s identity in National Tai-Chung Teachers Collage, the score of female is about 53.042 and male’s average score is 55.012.
4. By the degree, the outcome of the graduated students is highest, but it just has little difference from the other degree students. Each degree average score approaches total average score. This result appears that degree does not affect the degree of self-organization.
|
|---|
