4. Two different machine learning techniques are applied to a numeric data set and the two models built are Model 1 and Model 2. The classification test errors as paired sequences, using a 10-fold cross validation over 10 trials, are recorded as follows: Model 1: Model 2: 0.1 0.7 0.4 0.1 0.7 0.3 0.2 0.5 0.1 (a) Which model is better if we apply "voting" (if any)? (b) Which model is better than the other, if we compare the differences of their average errors, error (model 1)-error(model2), by using the paired Student's t-test with 9 degrees of freedom at the 10% level (i.e., 80% confidence level)? Compute and show work. 0.9 0.7 0.5 0.3 0.5 0.4 0.3 0.8 0.2 0.1 0.2

4. Two different machine learning techniques are applied to a numeric data set and the two models built are Model 1 and Model 2. The classification test errors as paired sequences, using a 10-fold cross validation over 10 trials, are recorded as follows: Model 1: Model 2: 0.1 0.7 0.4 0.1 0.7 0.3 0.2 0.5 0.1 (a) Which model is better if we apply "voting" (if any)? (b) Which model is better than the other, if we compare the differences of their average errors, error (model 1)-error(model2), by using the paired Student's t-test with 9 degrees of freedom at the 10% level (i.e., 80% confidence level)? Compute and show work. 0.9 0.7 0.5 0.3 0.5 0.4 0.3 0.8 0.2 0.1 0.2

Database System Concepts

7th Edition

ISBN:9780078022159

Author:Abraham Silberschatz Professor, Henry F. Korth, S. Sudarshan

Publisher:Abraham Silberschatz Professor, Henry F. Korth, S. Sudarshan

Chapter1: Introduction

Section: Chapter Questions

Problem 1PE

See similar textbooks

Similar questions

Draw the decision tree for training sample set in the picture with the CART (Classification and Regression Tree) algorithm by making the necessary calculations. Thank you so much.
3. Write a code to compute the 1%, 5% and 10% critical values for unit root test using the statistic: n(ô – 1) for sample sizes n = 50, 100, 200 and 1000 for the following two cases: Case 1. Generate Y; by Yt = Yt-1+ut, and estimate p based on Y = pYt-1 + ut. Case 2. Generate Y; by Y = Yt-1+ Ut, and estimate p based on Y = c+ pYt-1+Ut.
+
Question 48. Let us return to the Titanic data set. We now have learned several models and want to choose the best one. We used three different methods to validate these models: The training error rate (apparent error rate), the error rate on an external test set and the error rate estimated by a 10-fold cross validation. Training Error | Error on the test set | Cross Validation Error 0.18 Learner Decision Tree 0.22 0.21 Random Forest 0.01 0.10 0.12 1-Nearest-Neighbour 0.18 0.19 Which of the following statements are correct? a) 1-Nearest-Neighbour has a perfect training error and hence it should be used here. b) Random Forests outperforms both 1-Nearest-Neighbour and the Decision Tree in terms of prediction error. c) Not just in this case, but in general, Cross Validation is the better validation strategy and should always be preferred over the error on a single test set. d) Not just in this case, but in general, Decision Trees always perform worse than Random Forests.
Exercise 1 To practice this skill, the code below samples 100 times from the population of values 0,1 and adds up the results to create a new random variable S = sum of all sample valued. Reduce the code below to three lines by removing the use of the intermediate variable x: Code Start Over 1 values-c(0,1) 2 x-sample(values, 100, replace-TRUE) 3 phat-sum(x--"0")/100 Exercise 2 After you understand the above reduction, the next step is to wrap this random variable calculation inside the replicate function and repeat the result of this calculation some large number of times N. The basic format of this function is as follows: replicate (N, random variable calculation) The code below performs N = 5 replicates of the random variable which calculates p in n = 10 tosses of a fair coin and prints the result. Alter the code so that it performs N= 10 replicates of the random variable p, but with only n = 5 tosses of a fair coin. Code Run Code Start Over 1 values-c("H","T") 2 phat-replicate(5,…
Computer Science In the context of supervised learning algorithms, explain when we should use each of these models: Linear Regression, Logistic Regression, Decision Tree, Randon Forest, Regression Tree, Neural Network. Make a flowchart showing reasons for the decision among all of them based on information from the input data. Do something as detailed as possible, not just separating into classification and regression problems..
Class Su pastry Kol pastry Kol pastry Kol pastry Su pastry Su pastry ???? Butter (g) Milk (ml) Flour (g) Water(ml) хоо 200 200 250 200 100 300 200 150 100 300 200 хоо 50 350 300 275 150 325 хоо 275 225 275 275 175 хоо 150 400 (Learning outcome 2: will be able to relate Data Mining Models with each other.) 2. (These QUESTIONS are NOT True FALSE questions you have to state and discuss your answer. IF you just write TRUE / FALSE and not describe the REASONS your answer WONT be scored) We get rid of missing and inconsistent data with preprocess step called data editing, which is one of the data mining preprocessing step. а.
Short questions. Please choose the right choices for each question. There may be more than one correct choice. 1. A machine learning model usually has a few hyper-parameters to tune. Which of the two following settings are good for choosing hyper-parameters that can generalize well to new data: (a) Choose hyperparameters that work best on the all data (b) Choose hyperparameters that work best on the training set (c) Choose hyperparameters that work best on the validation set (d) Choose hyperparameters by performing k-fold cross-validation.
Use decision tree to further explore the dataset, where the dependent variable is ‘smoker’. Please explain the approach taken. [No more than 300 words]
The input required by any supervised learning algorithm is O a. Labeled dataset O b. Neither labeled nor unlabeled dataset O c. Labelled as well as unlabeled dataset O d. Unlabeled dataset
2. Can you design a binary classification experiment with 100 total population (TP+TN+FP+ FN), with precision (TP/(TP+FP)) of 1/2, with sensitivity (TP/(TP+FN)) of 2/3, and specificity (TN/(FP+TN)) of 3/5? (Please consider the population to consist of 100 individuals.)
Butter (g) Milk (ml) Flour (g) Water(ml) Class Su pastry Kol pastry Kol pastry Kol pastry Su pastry Su pastry ???? хо 200 250 200 200 100 300 200 150 100 300 200 хо 50 350 300 275 275 150 175 325 хоо X00 275 225 150 275 400 (Learning outcome 2: will be able to relate Data Mining Models with each other.) 2. (These QUESTIONS are NOT True FALSE questions you have to state and discuss your answer. IF you just write TRUE / FALSE and not describe the REASONS your answer WONT be scored)