Question

In: Math

Compare the kk-NN classifier, linear discriminant analysis (LDA) and the logistic model when it comes to...

Compare the kk-NN classifier, linear discriminant analysis (LDA) and the logistic model when it comes to classification. Which is generally better?

Solutions

Expert Solution

Answer:

In the LDA framework, we can see that the log odds is given by

where c0 and c1 are functions of µ1, µ2 and σ 2 .

By formulation of logistic regression, we can see that,

Both are linear functions of x. Hence, both logistic regression and LDA produce linear decision boundaries. The only difference between the two approaches lies in the fact that β0 and β1 are estimated using maximum likelihood, whereas c0 and c1 are computed using the estimated mean and variance from a normal distribution.

(  Since logistic regression and LDA differ only in their fitting procedures, one might expect the two approaches to give similar results. This is often, but not always, the case. LDA assumes that the observations are drawn from a Gaussian distribution with a common covariance matrix in each class, and so can provide some improvements over logistic regression when this assumption approximately holds. Conversely, logistic regression can outperform LDA if these Gaussian assumptions are not met.)

KNN

KNN takes a completely different approach from those classifiers. In order to make a prediction for an observation X = x, the K training observations that are closest to x are identified. Then X is assigned to the class to which the plurality of these observations belong. Hence KNN is a completely non-parametric approach: no assumptions are made about the shape of the decision boundary. Therefore, we can expect this approach to dominate LDA and logistic regression when the decision boundary is highly non-linear. On the other hand, KNN does not tell us which predictors are important; we don't get a table of coefficients.

In short which classifier is better is depend on situation.

For parametric gaussian data LDA is better. For parametric non-gaussian data Logistic regression is better. And for non-parametric data knn is better.


Related Solutions

How would you differentiate among multiple discriminant analysis, regression analysis, logistic regression analysis, and analysis of...
How would you differentiate among multiple discriminant analysis, regression analysis, logistic regression analysis, and analysis of variance and demonstrate statistical significance for each?
When should logistic regression be used for data analysis? What is the assumption of logistic regression?...
When should logistic regression be used for data analysis? What is the assumption of logistic regression? How to explain odds ratio?
Consider a logistic regression model with a linear predictor that includes an interaction term, say x’β...
Consider a logistic regression model with a linear predictor that includes an interaction term, say x’β = β0 + β1x1 +β2x2+ β12x1x2. Does the odds ratio associated with the regressor x2 have the same interpretation as in the case where the linear predictor does not have the interaction term? Provide mathematical development to support your answer.
1.When is logistic regression the appropriate model for modeling non-metric outcomes? 2.In what ways is logistic...
1.When is logistic regression the appropriate model for modeling non-metric outcomes? 2.In what ways is logistic regression comparable to multiple regression? How does it differ? 3.Why are there two forms of logistic coefficients (original and exponentiated)?
1) True or False? In multinomial logistic regression analysis, the model fit is given by the...
1) True or False? In multinomial logistic regression analysis, the model fit is given by the value of R-squared. In linear regression analysis, the overall significance test is an F test. In logistic regression analysis, the overall significance test is an F test. When building a model for regression analysis, the type of data for the outcome variable guides you to choose linear, logistic, or multinomial logistic. Two variables that covary might not have a causal relationship. Causation implies correlation....
What are the major differences between a static model and a dynamic model when it comes...
What are the major differences between a static model and a dynamic model when it comes to valuation of natural resources? In your opinion, which model is more suitable for valuating depletable natural resources? And why? What are the main factors that could affect the valuation of natural resources? Use example(s) if necessary.
When we estimate a linear multiple regression model (including a linear simple regression model), it appears...
When we estimate a linear multiple regression model (including a linear simple regression model), it appears that the calculation of the coefficient of determination, R2, for this model can be accomplished by using the squared sample correlation coefficient between the original values and the predicted values of the dependent variable of this model. Is this statement true? If yes, why? If not, why not? Please use either matrix algebra or algebra to support your reasoning.
T/F   In a logistic model when N is equal to K the population stops growing and...
T/F   In a logistic model when N is equal to K the population stops growing and R =0 In the logistic model of population growth if r>0 then Both the population decline and N&lt;K Population growth stops N<K N>K The population declines If the age distribution diagram has the smallest concentration of individuals in the bottom portion or is considered to be an urn shape, this indicates that Cannot be determined The population is growing slowly The population is declining...
Compare and contrast between the business analysis process model and the business analysis core concept model....
Compare and contrast between the business analysis process model and the business analysis core concept model. business analysis and financial policy
Discuss and explain the log-linear model and the logit model. Give examples of when these models...
Discuss and explain the log-linear model and the logit model. Give examples of when these models are used?
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT