This set of Neural Networks Multiple Choice Questions & Answers (MCQs) focuses on “Neural Networks – Non-Linear Hypothesis”.
1. The given non linear classification problem can be solved using linear methods.
a) True
b) False
View Answer
Explanation: The given figure shows a non linear classification of data points. Here the dark bubbles and dark triangles represent the data points which cannot be linearly separated using any linear methods. So we need some non linear classification model to learn a complex non-linear hypothesis for many applications like this.
2. The given classification problem cannot be well solved by using logistic regression.
a) True
b) False
View Answer
Explanation: Given a highly non-linear classification task and in order to achieve a decision boundary like the one plotted in the below figure, one needs to introduce non-linear features in the form of quadratic and other higher order terms. Hence logistic regression would not generalize the solution very well by leveraging the power of polynomial terms.
3. From the given figure as the number of features increase the number of terms in the hypotheses would also increase exponentially to get a good fit which would have high probability of overfitting the data.
a) True
b) False
View Answer
Explanation: Given a non linear classification problem. Here as the number of features increase the number of terms in the hypotheses would also increase exponentially to get a good fit which would have high probability of overfitting the data. Hence, when the number of features is really high and the decision boundary is complex we have to use neural networks to solve the problem.
4. Which of the following methods can be used to solve the below classification problem?
a) Linear regression
b) Multivariate linear regression
c) Simple logistic regression
d) Non linear logistic regression
View Answer
Explanation: Linear regression, Multivariate linear regression and simple logistic regression are not enough to solve the given non linear classification problem. We have to define a non-linear logistic regression as a hypothesis H. And our goal is to find a good H which can distinguish dark circle data and dark square data well.
Sanfoundry Global Education & Learning Series – Neural Networks.
To practice all areas of Neural Networks, here is complete set on 1000+ Multiple Choice Questions and Answers.
