1 . Classification Of Machine Learning Algorithms
Machine learning can be classified by many different yardsticks some of the most basic are :
Classification based on what type of problem the algorithm aims to solve
Type 1 : Classification type problems , This type of algorithm classifies the given data according to set rules into sets of data . For example a classification algorithm can be used to classify a given set of people into male and female depending on the their genetic data . So the classification problem groups data into clumps of internally related data groups ( male and female for the previous example ) . The number of groups can be decided by the relation by which data is to be classified .
For example applying the “ Species” rule on a dataset comprising of “pets” will result in the algorithm creating two or more resultant groups , namely “Cat” and “Dog ” .Similarly applying the “Race” rule on the set “Humans” will result in creation of resultant groups like “Mongoloid ”, “Caucasian” , “African” etc .
Type 2 : Regression or relationship estimation , This type of problem tries to establish a relation between the inputs fed to it . This algorithm attempts to create a mathematical relationship between the various inputs , thus mapping a relationship between them .This relationship can be used to predict the future values like in stock , bond or housing price prediction . The major difference with respect to the previous type (Classification) is that the output of such problems is continuous valued where as Classification give a discrete valued output .
Many techniques for carrying out regression analysis have been developed. Familiar methods such as linear regression…en.wikipedia.org
Continued in next post ……
