Skip to main content icon/video/no-internet

Human Classification Learning

Human classification learning is the process of learning to distinguish between different types or groups of things in the world. Humans are often presented with various situations in which people (friend or foe), objects (edible or toxic), and places (safe or dangerous) must be correctly classified to survive. Classification allows us to respond to these situations appropriately and further allows these situations to be identified and incorporated into groups, categories, or concepts based on similar properties. This entry will further define human classification learning as well as highlight different ways in which classification can be studied. Finally, the entry will briefly introduce the neural architecture underlying human classification learning.

To group or categorize various types of objects and situations, learners can either learn the group membership of each object or situation individually or can generalize by finding common attributes within a group of objects, with the aim of being able to apply the same action to these distinct, yet similar, objects or situations in future encounters. Classification learning can have several uses. Classification allows us to bypass the process of full identification of a stimulus, to quickly generalize across objects or situations in our environment on the basis of simple dimensions, and to respond appropriately to the environment. Classification allows us to quickly and efficiently distinguish among a wide range of objects even without the benefit of prior experience. In addition, classification enables us to easily group different objects so that more complex grouping structures within specific categories can be abstracted. In other words, classification is the initial process by which objects are parsed into separate groups before category membership can be determined, a necessary step one must go through to access information acquired through category learning.

Though similar, human classification learning should not be confused with category learning or even concept learning. Overall, classification emphasizes the processes involved in assigning stimuli to groups, whereas concept learning emphasizes the mental representations of our semantic knowledge of the world. Categorization often combines the process of classification with the mental representations of concept learning. However, at times these terms have been used interchangeably, so researchers and students in this area should take care when reading the scientific literature.

Experimental Approaches

Because of its wide scope, human classification learning can include any task that requires the learner to make distinctions between the group membership of two or more stimuli. Despite this, classification is typically studied experimentally using three approaches. One approach is known as trial and error learning, or feedback-based learning. In a typical study, the goal of a learner is to produce an appropriate response for any stimulus without the benefit of previous training examples. To achieve this, the learner must view the stimulus, then make an initial response, and finally receive feedback. Based on the resulting feedback, the learner must abstract sorting criteria and generalize these rules to new stimuli in some “reasonable” way. Eventually, this trial and error method will enable the learner to produce the correct response more and more frequently. A second approach is instructed learning, or paired-associate learning. Learners are initially told how each stimulus should be classified, with the goal of eventually being able to classify stimuli without being given the answer. A third approach is known as free sorting. This approach requires learners to partition stimuli along one or more dimensions without information about whether their response was correct or incorrect. Learning has to rely on the distinguishing attributes of the stimuli for the stimuli to be classified. Formally, trial and error learning and instructed learning are both types of supervised learning, whereas free sorting is a form of unsupervised learning; the theoretical differences between these two approaches are beyond the scope of this entry.

...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles

Sage Recommends

We found other relevant content for you on other Sage platforms.

Loading