ベストな準備プランDatabricks-Certified-Professional-Data-Scientist試験2022年最新のDatabricks Certification無制限140問題
注目すべき時短になるDatabricks-Certified-Professional-Data-Scientistオールインワン試験ガイド
Databricks Databricks-Certified-Professional-Data-Scientist 認定試験の出題範囲：
| トピック | 出題範囲 |
|---|---|
| トピック 1 |
|
| トピック 2 |
|
| トピック 3 |
|
| トピック 4 |
|
| トピック 5 |
|
| トピック 6 |
|
質問 76
Reducing the data from many features to a small number so that we can properly visualize it in two or three dimensions. It is done in_______
- A. k-Nearest Neighbors
- B. Support vector machines
- C. supervised learning
- D. un-supervised learning
正解: D
解説:
Explanation
The opposite of supervised learning is a set of tasks known as unsupervised learning. In unsupervised learning, there's no label or target value given for the data. A task where we group similar items together is known as clustering. In unsupervised learning, we may also want to find statistical values that describe the data. This is known as density estimation. Another task of unsupervised learning may be reducing the data from many features to a small number so that we can properly visualize it in two or three dimensions
質問 77
Which of the following true with regards to the K-Means clustering algorithm?
- A. Labels are pre-assigned to each objects in the cluster.
- B. Labels are not pre-assigned to each objects in the cluster.
- C. It discovers the center of each cluster.
- D. It find each objects fall in which particular cluster
- E. It classify the data based on the labels.
正解: B,C,D
解説:
Explanation
Clustering does not require any predefined labels on the object, rather it consider the attributes on the object.
Hence, option-B is out. Clustering is different than classification technique.
Hence you can discard the option-C as well. It does not use the pre-defined labels, hence it is called unsupervised learning and option-Ais correct. Main purpose of the Clustering technique is to determine the center of each Cluster and then find the distance from that center. If object is near the center than it would fall in that particular cluster. Hence, finally you will have group or clusters created and get to know that objects fall in which particular cluster.
質問 78
Select the correct option which applies to L2 regularization
- A. Non-sparse outputs
- B. Computational efficient due to having analytical solutions
- C. No feature selection
正解: A,B,C
解説:
The difference between their properties can be promptly summarized as follows:
Table Description automatically generated
質問 79
Which method is used to solve for coefficients bO, b1, ... bn in your linear regression model:
- A. Integer programming
- B. Ridge and Lasso
- C. Apriori Algorithm
- D. Ordinary Least squares
正解: D
解説:
Explanation : RY = b0 + b1x1+b2x2+ .... +bnxn
In the linear model, the bi's represent the unknown p parameters. The estimates for these unknown parameters are chosen so that, on average, the model provides a reasonable estimate of a person's income based on age and education. In other words, the fitted model should minimize the overall error between the linear model and the actual observations. Ordinary Least Squares (OLS) is a common technique to estimate the parameters
質問 80
Feature Hashing approach is "SGD-based classifiers avoid the need to predetermine vector size by simply picking a reasonable size and shoehorning the training data into vectors of that size" now with large vectors or with multiple locations per feature in Feature hashing?
- A. It is hard to understand what classifier is doing
- B. Is a problem with accuracy as well as hard to understand what classifier us doing
- C. It is easy to understand what classifier is doing
- D. Is a problem with accuracy
正解: A
解説:
Explanation
FEATURE HASHING
SGD-based classifiers avoid the need to predetermine vector size by simply picking a reasonable size and shoehorning the training data into vectors of that size. This approach is known as feature hashing. The shoehorning is done by picking one or more locations by using a hash of the name of the variable for continuous variables or a hash of the variable name and the category name or word for categorical, text*like, or word-like data.
This hashed feature approach has the distinct advantage of requiring less memory and one less pass through the training data, but it can make it much harder to reverse engineer vectors to determine which original feature mapped to a vector location. This is because multiple features may hash to the same location. With large vectors or with multiple locations per feature, this isn't a problem for accuracy but it can make it hard to understand what a classifier is doing.
An additional benefit of feature hashing is that the unknown and unbounded vocabularies typical of word-like variables aren't a problem.
質問 81
You are working on a Data Science project and during the project you have been gibe a responsibility to interview all the stakeholders in the project. In which phase of the project you are?
- A. Data Preparations
- B. Creating Models
- C. Operationnalise the models
- D. Creating visuals from the outcome
- E. Executing Models
- F. Discovery
正解: F
解説:
Explanation
During the discovery phase you will be interviewing all the project stakeholders because they would be having quite a good amount of knowledge for the problem domain you will be working and you also interviewing project sponsors you will get to know what all are the expectations once project get completed. Hence, you will be noting down all the expectations from the project as well as you will be using their expertise in the domain.
質問 82
You are working as a data science consultant for a gaming company. You have three member team and all other stake holders are from the company itself like project managers and project sponsored, data team etc.
During the discussion project managed asked you that when can you tell me that the model you are using is robust enough, after which step you can consider answer for this question?
- A. Model planning
- B. Model building
- C. Discovery
- D. Data Preparation
- E. Operationalize
正解: B
解説:
Explanation
To answer whether the model you are building is robust enough or not you need to have answer below questions at least
- Model is performing as expected with the test data or not?
- Whatever hypothesis defined in the initial phase is being tested or not?
- Do we need more data?
- Domain experts are convinced or not with the model?
And all these can be answered when you have built the model and tested with the test data sets. Hence, correct option will be Model Building.
質問 83
You are doing advanced analytics for the one of the medical application using the regression and you have two variables which are weight and height and they are very important input variables, which cannot be ignored and they are also highly co-related. What is the best solution for that?
- A. You would consider using BMI (Body Mass Index)
- B. You will take square of the height.
- C. You will take square root of weight
- D. You will take cube root of height
正解: A
解説:
Explanation
If multiple variables are highly co-related then it is better you consider using the either of the variable which correlates more (which is not in the given option) or go for the new variable which is a function of the both the variable in this case it could be BMI (Body Mass Index). Because it is a function of both weight and height as per the below formula. BMI = Weight/(Height * Height)
質問 84
What describes a true property of Logistic Regression method?
- A. It is robust with redundant variables and correlated variables.
- B. It handles missing values well.
- C. It works well with discrete variables that have many distinct values.
- D. It works well with variables that affect the outcome in a discontinuous way.
正解: A
質問 85
Assume some output variable "y" is a linear combination of some independent input variables "A" plus some independent noise "e". The way the independent variables are combined is defined by a parameter vector B y=AB+e where X is an m x n matrix. B is a vector of n unknowns, and b is a vector of m values. Assuming that m is not equal to n and the columns of X are linearly independent, which expression correctly solves for B?
- A. Option D
- B. Option B
- C. Option A
- D. Option C
正解: A
解説:
Explanation
This is the standard solution of the normal equations for linear regression. Because A is not square, you cannot simply take its inverse.
質問 86
In which of the following scenario you should apply the Bay's Theorem
- A. Within the sample space, there exists an event B, for which P(B) > 0.
- B. The analytical goal is to compute a conditional probability of the form: P(Ak | B ).
- C. In all above cases
- D. The sample space is partitioned into a set of mutually exclusive events {A1, A2, . .., An }.
正解: C
質問 87
Suppose there are three events then which formula must always be equal to P(E1|E2,E3)?
- A. P(E1,E2,E3)P(E1)/P(E2:E3)
- B. P(E1,E2;E3)/P(E2,E3)
- C. P(E1,E2|E3)P(E3)
- D. P(E1,E2|E3)P(E2|E3)P(E3)
- E. P(E1,E2,E3)P(E2)P(E3)
正解: B
解説:
Explanation
This is an application of conditional probability: P(E1,E2)=P(E1|E2)P(E2). so P(E1|E2) = P(E1.E2)/P(E2) P(E1,E2,E3)/P(E2,E3) If the events are A and B respectively, this is said to be "the probability of A given B" It is commonly denoted by P(A|B): or sometimes PB(A). In case that both "A" and "B" are categorical variables, conditional probability table is typically used to represent the conditional probability.
質問 88
As a data scientist consultant at ABC Corp, you are working on a recommendation engine for the learning resources for end user. So Which recommender system technique benefits most from additional user preference data?
- A. Content-based filtering
- B. Logistic Regression
- C. Naive Bayes classifier
- D. Item-based collaborative filtering
正解: D
解説:
Explanation
Item-based scales with the number of items, and user-based scales with the number of users you have. If you have something like a store, you'll have a few thousand items at the most. The biggest stores at the time of writing have around 100,000 items. In the Netflix competition, there were 480,000 users and 17,700 movies. If you have a lot of users: then you'll probably want to go with item-based similarity. For most product-driven recommendation engines, the number of users outnumbers the number of items. There are more people buying items than unique items for sale. Item-based collaborative filtering makes predictions based on users preferences for items. More preference data should be beneficial to this type of algorithm. Content-based filtering recommender systems use information about items or users, and not user preferences, to make recommendations. Logistic Regression, Power iteration and a Naive Bayes classifier are not recommender system techniques.
質問 89
Which of the following is a Continuous Probability Distributions?
- A. Poisson probability distribution
- B. Binomial probability distribution
- C. Negative binomial distribution
- D. Normal probability distribution
正解: D
質問 90
Select the correct statement which applies to Principal component analysis (PCA)
- A. Increase the dimensionality of the data set.
- B. Is a mathematical procedure that transforms a number of (possibly) correlated variables into a (higher) number of uncorrelated variables
- C. Is a mathematical procedure that transforms a number of (possibly) correlated variables into a (smaller) number of uncorrelated variables.
- D. 1 and 3 are correct
- E. 1 and 2 are correct
正解: C
解説:
Explanation
Principal component analysis (PCA) involves a mathematical procedure that transforms a number of (possibly) correlated variables into a (smaller) number of uncorrected variables called principal components. The first principal component accounts for as much of the variability in the data as possible, and each succeeding component accounts for as much of the remaining variability as possible.
質問 91
Your company has organized an online campaign for feedback on product quality and you have all the responses for the product reviews, in the response form people have check box as well as text field. Now you know that people who do not fill in or write non-dictionary word in the text field are not considered valid feedback. People who fill in text field with proper English words are considered valid response. Which of the following method you should not use to identify whether the response is valid or not?
- A. Random Decision Forests
- B. Logistic Regression
- C. Naive Bayes
- D. Any one of the above
正解: D
解説:
Explanation
In this problem you have been given high-dimensional independent variables like yeS; nO; no English words , test results etc. and you have to predict either valid or not valid (One of two). So all of the below technique can be applied to this problem.
* Support vector machines
* Naive Bayes
* Logistic regression
* Random decision forests
質問 92
A bio-scientist is working on the analysis of the cancer cells. To identify whether the cell is cancerous or not, there has been hundreds of tests are done with small variations to say yes to the problem. Given the test result for a sample of healthy and cancerous cells, which of the following technique you will use to determine whether a cell is healthy?
- A. Identification Test
- B. Naive Bayes
- C. Linear regression
- D. Collaborative filtering
正解: B
解説:
Explanation
In this problem you have been given high-dimensional independent variables like yes, no: test results etc. and you have to predict either valid or not valid (One of two). So all of the below technique can be applied to this problem.
Support vector machines Naive Bayes Logistic regression Random decision forests
質問 93
Which of the following are advantages of the Support Vector machines?
- A. SVMs directly provide probability estimates
- B. it is memory efficient
- C. Number of features is much greater than the number of samples, the method still give good performances
- D. Effective in high dimensional spaces.
- E. Effective in cases where number of dimensions is greater than the number of samples
- F. possible to specify custom kernels
正解: B,D,E,F
解説:
Explanation
Support vector machines (SVMs) are a set of supervised learning methods used for classification, regression and outliers detection.
The advantages of support vector machines are:
Effective in high dimensional spaces.
Still effective in cases where number of dimensions is greater than the number of samples.
Uses a subset of training points in the decision function (called support vectors), so it is also memory efficient.
Versatile: different Kernel functions can be specified for the decision function.
Common kernels are provided, but it is also possible to specify custom kernels.
The disadvantages of support vector machines include:
If the number of features is much greater than the number of samples, the method is likely to give poor performances.
SVMs do not directly provide probability estimates, these are calculated using an expensive five-fold cross-validation.
質問 94
You are building a classifier off of a very high-dimensiona data set similar to shown in the image with 5000 variables (lots of columns, not that many rows). It can handle both dense and sparse input. Which technique is most suitable, and why?
- A. Naive Bayes, because Bayesian methods act as regularlizers
- B. Logistic regression with L1 regularization, to prevent overfitting
- C. k-nearest neighbors, because it uses local neighborhoods to classify examples
- D. Random forest because it is an ensemble method
正解: B
解説:
Explanation
Logistic regression is widely used in machine learning for classification problems. It is well-known that regularization is required to avoid over-fitting, especially when there is a only small number of training examples, or when there are a large number of parameters to be learned. In particular L1 regularized logistic regression is often used for feature selection, and has been shown to have good generalization performance in the presence of many irrelevant features. (Ng 2004; Goodman 2004) Unregularized logistic regression is an unconstrained convex optimization problem with a continuously differentiate objective function. As a consequence, it can be solved fairly efficiently with standard convex optimization methods, such as Newton's method or conjugate gradient. However, adding the L1 regularization makes the optimization problem com-putationally more expensive to solve. If the L1 regulariza-tion is enforced by an L1 norm constraint on the parameLogistic regression is a classifier and L1 regularization tends to produce models that ignore dimensions of the input that are not predictive. This is particularly useful when the input contains many dimensions, k-nearest neighbors classification is also a classification technique, but relies on notions of distance. In a high-dimensional space, most every data point is "far" from others (the curse of dimensionality) and so these techniques break down. Naive Bayes is not inherently regularizing. Random forests represent an ensemble method; but an ensemble method is not necessarily more suitable to high-dimensional data.
Practically, I think the biggest reasons for regularization are 1) to avoid overfitting by not generating high coefficients for predictors that are sparse. 2) to stabilize the estimates especially when there's collinearity in the data.
1) is inherent in the regularization framework. Since there are two forces pulling each other in the objective function, if there's no meaningful loss reduction, the increased penalty from the regularization term wouldn't improve the overall objective function. This is a great property since a lot of noise would be automatically filtered out from the model. To give you an example for 2), if you have two predictors that have same values, if you just run a regression algorithm on it since the data matrix is singular your beta coefficients will be Inf if you try to do a straight matrix inversion. But if you add a very small regularization lambda to it, you will get stable beta coefficients with the coefficient values evenly divided between the equivalent two variables. For the difference between L1 and L2, the following graph demonstrates why people bother to have L1 since L2 has such an elegant analytical solution and is so computationally straightforward. Regularized regression can also be represented as a constrained regression problem (since they are Lagrangian equivalent). The implication of this is that the L1 regularization gives you sparse estimates. Namely, in a high dimensional space, you got mostly zeros and a small number of non-zero coefficients. This is huge since it incorporates variable selection to the modeling problem. In addition, if you have to score a large sample with your model, you can have a lot of computational savings since you don't have to compute features(predictors) whose coefficient is 0. I personally think L1 regularization is one of the most beautiful things in machine learning and convex optimization. It is indeed widely used in bioinformatics and large scale machine learning for companies like Facebook, Yahoo, Google and Microsoft.
質問 95
Select the correct option from the below
- A. If you're trying to predict or forecast a target value^ then you need to look into supervised learning.
- B. If you're not trying to predict a target value, then you need to look into unsupervised learning
- C. Are you trying to fit your data into some discrete groups? If so and that's all you need, you should look into clustering.
- D. If the target value can take on a number of values, say any value from 0.00 to 100.00, or -999 to 999: or
+_to -_, then you need to look unsupervised learning - E. If you've chosen supervised learning, with discrete target value like Yes/No. 1/2/3, A/B/C: or Red/Yellow/Black, then look into classification.
正解: A,B,C,E
解説:
Explanation
If you re trying to predict or forecast a target value, then you need to look into supervised learning. If not, then unsupervised learning is the place you want to be. If you've chosen supervised learning, what's your target value? Is it a discrete value like Yes/No, 1/2/3, A/B/C: or Red/Yellow/Black? If so, then you want to look into classification. If the target value can take on a number of values, say any value from 0.00 to 100.00, or-999 to
999, or+_to -_, then you need to look into regression. If you're not trying to predict a target value: then you need to look into unsupervised learning. Are you trying to fit your data into some discrete groups? If so and that's all you need, you should look into clustering. Do you need to have some numerical estimate of how strong the fit is into each group? If you answer yes then you probably should look into a density estimation algorithm.
質問 96
......
合格保証付きDatabricks-Certified-Professional-Data-Scientist問題集：https://jp.fast2test.com/Databricks-Certified-Professional-Data-Scientist-premium-file.html