类别学习中的典型性效应和诊断性效应
发布时间：2018-07-17 08:24
【摘要】：类别学习是人类认识世界的一项基本认知活动,类别学习能力是人类的基本能力。类别学习主要存在两种学习方式:分类学习和推理学习。分类学习是指告知类别所有信息,判断类别标签;推理学习是告知类别标签和部分特征值,判断缺失特征值。当前研究所使用的的类别材料多是光栅图和家族相似性图片材料,更多的进行的是知觉加工,因此也称之为知觉类别学习。但是在现实生活中,自然界的刺激并不单单是以图像形式呈现的,还有相当一部分是以言语或文字的形式呈现,这种呈现方式更多的涉及到语义加工,因此也称之为语义类别学习。一方面,在以往研究中,分类学习和推理学习对知觉类别学习的影响还存在争议,哪种学习方式更有利于知觉类别学习,并且学习的结果是怎样的有待更多的证据支持;另一方面,目前对语义类别学习的研究相对比较少,那么在语义类别学习中,分类学习和推理学习的影响又是怎样的尚不清晰,如果有影响,是怎样影响的。因此本研究主要通过两个实验探讨知觉类别学习和语义类别学习的影响机制。为了是两者更具有比较性,语义类别学习的学习材料是直接把知觉类别学习的材料用文字表述,这样可以在一定程度上保证两种学习材料在难度上和传达的信息等方面尽可能相同。实验一主要通过两个小实验分别探讨在知觉学习条件下和语义类别学习条件下,分类学习和推理学习的结果。结果表明,在知觉类别学习条件下,推理学习优于分类学习;在语义类别学习条件下,分类学习优于推理学习,知觉类别学习和语义类别学习产生了相反的结果,这说明知觉类别学习和语义类别学习可能存在不同的影响机制;实验二则在实验一的基础上进一步探讨知觉类别学习和语义类别学习的影响机制。通过典型性评价和单特征测试考察在两种学习条件下的特征典型性效应和样例诊断性效应。通过上述两个研究,主要得出如下结论:在知觉类别学习中,推理学习优于分类学习;在语义类别学习条件下,分类学习优于推理学习,主要体现在学习成绩和策略转换速度;知觉类别学习和语义类别学习都会受到样例典型性的影响。知觉类别学习和语义类别学习都能进行原型学习;知觉类别学习和语义类别学习都会存在诊断性效应。在知觉类别学习中,诊断性特征对分类学习很重要,而在语义类别学习中,诊断性特征的学习成绩更好,但是对非诊断性特征的学习,分类学习好于推理学习。说明在语义类别学习中,分类学习的注意广度更广泛。
[Abstract]:Category learning is a basic cognitive activity for human beings to understand the world, and category learning ability is the basic ability of human beings. There are two main types of learning in category learning: classified learning and reasoning learning. Classification learning means to tell all the information of the category and judge the label of the category, while the reasoning learning is to tell the label of the category and some characteristic values, and to judge the missing feature value. At present, most of the materials used in the research are raster images and family similarity images, and most of them are perceptual processing, so they are also called perceptual class learning. But in real life, the stimulation of nature is not only presented in the form of images, but also in the form of words or words, which is more involved in semantic processing. Therefore, it is also called semantic category learning. On the one hand, the influence of classification learning and reasoning learning on perceptual category learning is still controversial, which learning method is more favorable to perceptual category learning, and what kind of learning results need more evidence to support; On the other hand, the current research on semantic category learning is relatively few, so in the semantic category learning, the impact of classification learning and reasoning learning is not clear, if there is an impact, how to influence. Therefore, this study mainly discusses the influence mechanism of perceptual class learning and semantic category learning through two experiments. In order to be more comparative, the learning material of semantic category learning is to express the material of perceptual category learning directly in words. This ensures that the two learning materials are as difficult as possible in terms of difficulty and information. Experiment one mainly discusses the results of classified learning and inference learning under the condition of perceptual learning and semantic category learning through two small experiments. The results show that under the condition of perceptual category learning, inference learning is superior to classification learning, and in semantic category learning condition, classification learning is superior to reasoning learning, and perceptual category learning and semantic category learning have the opposite results. This shows that there may be different influence mechanisms between perceptual category learning and semantic category learning, while experiment two further discusses the influence mechanism of perceptual category learning and semantic category learning on the basis of experiment 1. Through the typical evaluation and the single feature test, the characteristic typicality effect and the sample diagnostic effect under the two learning conditions were investigated. Through the above two studies, the following conclusions are drawn: in perceptual category learning, reasoning learning is superior to classification learning; in semantic category learning, classification learning is superior to reasoning learning, which is mainly reflected in learning achievement and strategy conversion speed. Both perceptual class learning and semantic class learning are affected by the typicality of the sample. Both perceptual category learning and semantic category learning can carry out prototype learning, and both perceptual category learning and semantic category learning have diagnostic effects. In perceptual category learning, diagnostic features are very important for classification learning, while in semantic category learning, diagnostic features have better learning performance, but classification learning is better than reasoning learning for non-diagnostic features. It shows that in semantic category learning, the attention span of classification learning is more extensive.
【学位授予单位】：广州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：B842.2
本文编号：2129790
[Abstract]:Category learning is a basic cognitive activity for human beings to understand the world, and category learning ability is the basic ability of human beings. There are two main types of learning in category learning: classified learning and reasoning learning. Classification learning means to tell all the information of the category and judge the label of the category, while the reasoning learning is to tell the label of the category and some characteristic values, and to judge the missing feature value. At present, most of the materials used in the research are raster images and family similarity images, and most of them are perceptual processing, so they are also called perceptual class learning. But in real life, the stimulation of nature is not only presented in the form of images, but also in the form of words or words, which is more involved in semantic processing. Therefore, it is also called semantic category learning. On the one hand, the influence of classification learning and reasoning learning on perceptual category learning is still controversial, which learning method is more favorable to perceptual category learning, and what kind of learning results need more evidence to support; On the other hand, the current research on semantic category learning is relatively few, so in the semantic category learning, the impact of classification learning and reasoning learning is not clear, if there is an impact, how to influence. Therefore, this study mainly discusses the influence mechanism of perceptual class learning and semantic category learning through two experiments. In order to be more comparative, the learning material of semantic category learning is to express the material of perceptual category learning directly in words. This ensures that the two learning materials are as difficult as possible in terms of difficulty and information. Experiment one mainly discusses the results of classified learning and inference learning under the condition of perceptual learning and semantic category learning through two small experiments. The results show that under the condition of perceptual category learning, inference learning is superior to classification learning, and in semantic category learning condition, classification learning is superior to reasoning learning, and perceptual category learning and semantic category learning have the opposite results. This shows that there may be different influence mechanisms between perceptual category learning and semantic category learning, while experiment two further discusses the influence mechanism of perceptual category learning and semantic category learning on the basis of experiment 1. Through the typical evaluation and the single feature test, the characteristic typicality effect and the sample diagnostic effect under the two learning conditions were investigated. Through the above two studies, the following conclusions are drawn: in perceptual category learning, reasoning learning is superior to classification learning; in semantic category learning, classification learning is superior to reasoning learning, which is mainly reflected in learning achievement and strategy conversion speed. Both perceptual class learning and semantic class learning are affected by the typicality of the sample. Both perceptual category learning and semantic category learning can carry out prototype learning, and both perceptual category learning and semantic category learning have diagnostic effects. In perceptual category learning, diagnostic features are very important for classification learning, while in semantic category learning, diagnostic features have better learning performance, but classification learning is better than reasoning learning for non-diagnostic features. It shows that in semantic category learning, the attention span of classification learning is more extensive.
【学位授予单位】：广州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：B842.2
【参考文献】
相关期刊论文 前10条
1 邢强;孙海龙;刘凯;夏静静;;非监控类别学习的分类策略与表征[J];心理科学进展;2015年08期
2 孙海龙;邢强;;反馈对知觉类别学习的影响及其认知神经生理机制[J];心理科学进展;2014年01期
3 陈琳;莫雷;郑允佳;王雨函;;类别学习的阻碍效应与双机制理论[J];心理科学;2013年02期
4 刘凤英;姚志刚;李红;;类别标签与典型性程度对类别特征推理的影响[J];心理科学;2011年05期
5 张恒超;阴国恩;;学习方式对关系类别间接性学习的影响[J];心理与行为研究;2010年04期
6 方燕红;张积家;;汉字词和图片命名与分类的比较[J];心理学报;2009年02期
7 刘志雅;莫雷;;两种学习模式下类别学习的结果:原型和样例[J];心理学报;2009年01期
8 温红博;郭永兴;莫雷;;刺激材料的整体性知觉对类别建构的影响[J];心理学报;2008年05期
9 刘凤英;张庆林;李红;姚志刚;;类别学习中分类学习与推理学习的对比研究[J];心理科学;2008年02期
10 刘志雅;莫雷;;类别学习中两种学习模式的比较研究:分类学习与推理学习[J];心理学报;2006年06期
，本文编号：2129790
本文链接：https://www.wllwen.com/shekelunwen/xinlixingwei/2129790.html
