Data Science Quiz
English
Elementary
Test your knowledge on Data Science. A must attempt quiz for everyone.
Description
32
Questions
30 sec
Per question
4:10
Average time
3.0
Contest Score
35
Participants
7 comments
Dark Dodo
Good quiz, but

Q3: If the data is not available then what is done in order to collect data? - Various experiments are carried out

Partially disagree. I can successfully use web scrapping techniques as well as query languages like SQL to retrieve data (if I have an access). After all, the goal is to get data. In theory, any means are suitable. Various experiments are correct, but this is only 1 of the ways.

Q19: Statistical Modelling assumes that data came from simple model. 

I do not agree with this. Why I cannot use nonlinear models models, for example, in Statistical Modelling? If so then which Modelling should I use instead?

Q31: Which of the following machine learning methods is based on artificial neural networks? - Deep Learning

Why decision trees cannot be based on artificial neural networks as well as logistic regression?
I think there should be multiple options.
Link: https://arxiv.org/abs/1806.06988
Link: https://link.springer.com/chapter/10.1007/978-1-4612-2018-3_10
Calm Quokka author
Thank you for your comments. I really appreciate it.

Here is my response in the same order of your doubt :

1. This must be Q4 not Q3. (I didn't enabled shuffling, so I can't be wrong in this)

Web scraping is used when data is available on internet (websites) . Query languages like SQL are used when data is available in database. When data is not available then various experiments are carried out to collect data.

So there is no issue here.
=========================================

2. Statistical modelling assumes that data come from simple model which includes linear model as well as some non linear models like quadratic etc but not complex models like x^(25), log(sin(cos(xlog(x)))), etc. To model complex relations we use algorithmic modelling which is also called as machine learning.

So there is no issue here too.
=========================================

3. This is the only one for which u can counter me. But here is my say :
a) Deep Learning is based on artificial neural networks. Logistic regression and decision trees methods are traditionally, conventionally and generally not based on neural networks although their neural network implementation can be found.
b) Also for this elementary level quiz, Deep Learning shall suffice for this answer. If u wanna go in depth too then most suitable answer which is Deep Learning could be selected.

And thank you once again for your time :)

=========================================
Dark Dodo
Calm Quokka thank you for the explanation
Calm Quokka author
Dark Dodo Thank You too :)
Able Crow
Good quiz overall, i have few comments about some questions:
- Q3: i don't completely agree with the notion of ML and deep learning being subsets of data science; for sure they are subsets of AI, but i rather think data science as a field somehow orthogonal to those that would leverage AI, ML and DL
- Q19: if you mean by "simple model" that the assumptions of the statistical models are "simple to understand" or well defined, then i would tend to agree; but the current formulation of the question is misleading to me
- Q30: i also consider scatter plot as a way to detect outliers
Calm Quokka author
Hi, thank you for your precious comments. Here is my response in order of your doubts -

1. I apologize for this :(
But I want to defend myself a bit by saying that in such case, most suitable answer which is Artificial Intelligence could be selected.
=========================================

2. Statistical modelling assumes that data come from simple model which includes linear model as well as some non linear models like quadratic etc but not complex models like x^(25), log(sin(cos(xlog(x)))), etc. To model complex relations we use algorithmic modelling which is also called as machine learning.

So I found no issue here.
=========================================

3. I want to apologize for this too.
But I want to defend myself by saying that in such case, most suitable answer which is [option d : both a and b i.e violin and box plot] could be selected as scatter plots are generally used for finding relationships while box and violin plots are used for detecting outliers.

Thank You very much once again :)

=========================================
Calm Quokka author
Hi Telegram quiz evaluaters :) , I am the creator of this 'Data Science Quiz'. My quiz has following three issues as pointed above by humble competitors or myself, which I summarise as below -

Q3 - Actual answer should be option d- all of the above

Q30 - Given answer is option d- both a and b. But actually option c is also among the correct answer.

Q31 - Question could be refined by adding the word "traditionally" as shown below :
Which of the following machine learning methods is traditionally based on artificial neural networks?

Thank You!!
Take the quiz to leave a comment