Machine Learning - Test
Uploaded by
ScribdTranslationsMachine Learning - Test
Uploaded by
ScribdTranslations1)
The Naive Bayes Algorithm was created with its structure based on Bayes' Theorem,
what does the following say:
Consider the following situation:
P1(1 man | 2 men - P(1 men | 2 men) - P(2 men)
P(2homen)
P1(2 men | 1 man - P(1 man | 2 men) - P(2 men)
_________________________________________
P(1man)
2) The figure below contains a pseudo-algorithm of the training stage of a method
from machine learning.
To more distant neighbors.
Naive Bayes.
Linear regression.
Multivariate regression.
To the closest neighbors.
3)
Consider the following case in which a recommendation system is used:
A movie streaming website adds labels to each of the movies in its database.
data. Every time a user logs in, the system identifies the
user characteristics and, based on labels, makes recommendations for new movies to
they will be assisted.
Select the option that contains which recommendation system method is
used in the case described in the excerpt above.
Recommendation based on interaction.
Recommendation of collaborative filtering type based on deep learning.
Content-based recommendation.
4)
Consider the following statements about ensemble algorithms:
A. Individual classifiers may be 'more certain' about a specific part.
of the function's domain.
B. Classifiers that are 'more accurate' may have their vote weighted more.
C. In most cases, clusters perform better than classifiers
individuals.
Select the alternative that contains the CORRECT statements:
B - C
A – C.
A – B – C.
5)
Carefully analyze the following statement:
This algorithm uses predictions from many random individuals, each of which may
"has a different variance, and groups all of them."
Select the alternative that represents the algorithm to which the above statement refers.
refer.
Neural network.
Random forest.
6)
João administered a multiple-choice test containing 120 questions for a large
group of people, and realized that there was a great variation in
distribution of grades.
João then decided to use an algorithm to group the students, so that each
the student could stay in the most suitable group according to their grade. In this way,
he will be able to prepare specific work for each group.
Select the option that contains the most suitable algorithm for performing this.
grouping.
Linear regression.
K-means.
7)
The following graphs illustrate a linear regression algorithm applied to four
different cases of data dispersion. For all cases, there is the same line of
linear regression y = 4 + 0.5x, regardless of what the statistics are
sufficient, such as: linear and angular coefficients; sum of the squares of the residues and
from regression; and number of observations.
Based on the graphs, analyze the statements below:
I. Graph B indicates that the data is suitable for the use of regression
of the linear type.
II. The graph C illustrates the possible existence of an outlier observation in
linear regression.
III. The graph D illustrates the possible existence of an outlier observation in
linear regression.
Select the alternative that correctly assesses these statements:
Only I and III are correct.
Only II and III are correct.
Only I is correct.
Only II is correct.
8)
A type of methodology for machine learning is predictive AM based on
probabilities.
Regarding the probabilistic methods of AM, judge the following items as
true (T) or false (F):
Probabilistic methods are used when there is incomplete information about
the data to be used.
Probabilistic methods use prior probabilities and likelihood.
between events in order to make predictions.
The most popular algorithm that is based on probabilistic methods is the Naive.
Bayes, also known as Naive Bayes or simple Bayes.
Classify the statements as true (T) and false (F) and mark the alternative.
that contains the correct sequence, respectively, from top to bottom.
V – F – V.
V – V – V.
9)
There are numerous machine learning algorithms, and each is more suitable
for a certain type of data. There is an algorithm that is usually sensitive
regarding the position of the initial seeds, which can lead to convergence
local. This algorithm is also sensitive to noise, as it takes into account
the deviations of the objects in relation to the centroid of each calculated cluster. By
lastly, this algorithm is not suitable for cases where the clusters have shape
non-convex.
Select the option that contains the correct name of the algorithm described by the text
above.
Support vector machine.
K-means.
10)
Predictive machine learning (ML) follows the paradigm of learning
supervised.
About predictive and supervised AM, analyze the following items:
I. Predictive AM aims to find a function from objects of
training, to then predict the label of any new data provided.
II. The AM clustering is a classic method of supervised AM, as it only ...
receives the data to be analyzed and groups it according to common patterns
identified.
III. Linear regression, whether it has 2 or n variables, is a technique that uses
data to create a linear mathematical function that best represents
possible, the data. Since it creates a linear function, one can say that, once created
the function, the algorithm is of the unsupervised type.
Select the option that indicates which statements are correct.
Only I and III.
Only I.
