Data Science For Everyone Final
questions with correct answers.
False:
establishing association is easy as long as two variables generally move together but that does not mean
they cause one another and there for have a causal link. ANS - Is the Following True or False:
Establishing causality is generally easier than establishing association
True:
the point of randomization is to make the two groups receiving treatment and caontrol as similar as
possible ANS - Is the Following True or False: Randomizing treatment and control reduces the risk of
potential confounders
False:
I am not 100% sure about this, but maybe try to think about it like smoking and cancer. smoking and
cancer are correlated, but smoking is not a sufficient cause to get it. ANS - Is the Following True or False:
if x is not a sufficient condition for y then x is not correlated with y.
False.
In the case of a list, multiplying L*2 will only result in the list being repeated twice. ANS - Is the
Following True or False: Suppose L is a list in python containing only numeric values, then L*2 will return
all the numeric values doubled.
false:
there are a lot of mistakes that can happen. Like if you are randomly selecting and you just take outliers
it will not equal the population parameter. ANS - True or False: A statistic calculated from a random
sample must be equal to the population peramiter
True:
yes that is by detention the idea of a test stat. ANS - Is the Following True or False:
The test statistic tells us the relative plausibility of the null vs alternative hypothesis.
,true ANS - Is the Following True or False
if we fail to reject the null hypothesis when the null is false we have committed a type 2 error
false:
empirical is what the distribution of what we actually observe.
sample is the distribution of what our sample predicts.
so if we roll a die 3 times that would be an empirical
if we make a computer assimilate rolling a die 3 times that would be a sampling ANS - Is the Following
True or False:
the sampling distribution is the empirical distribution of given samples values.
true:
This is what a bootstrap is ANS - Is the Following True or False:
when calculating the sample distribution for a bootstrap we use replacement with our observations
false:
95 out of 100 will be in it, but any given sample is binary so it does not have a probability. ANS - Is the
Following True or False:
a given 95% confidence interval captures the true value of a parameter with a probability of 95%
true
yeah as you add more and more data points things tend to bunch away from outliers. ANS - Is the
Following True or False:
as our sample size increases the standard dev of the sample distribution is smaller
false there are decreasing marginal returns ANS - Is the Following True or False:
The standard error decreases in a linear fashion ie increasing sample size by 2 decrease error by 2.
false ANS - Is the Following True or False:
the data used in data science is always numeric
, true
that is a feature ANS - Is the Following True or False
correlation r is unit less.
true. ANS - Is the Following True or False:
a given 90% confidence interval captures the true value of the parameter or it does not
the basic idea of association is that it is a measure of how related two variables are. it is unit less
because it just a measure of how one thing compares to another. ANS - association
in order to standardize association, we take points and subtract the mean form them and then decide
them by the standard deviation to put everything in a way that is unit less and standard and in terms of
how many standard deviations they are form the mean. also z score is centerd at 0 , and 1 is 1 stnadrd
deviationa way. this allows us to compare data ANS - how does association relate to z scores
correctional coefficient is written as r, it is a measure of how strong an association is and given between
{-1,1} association is about how two variables are related, correlation is about how two variables are
associated ANS - what is correlation, correlation coefficient and how do they relate to linearity.
it is using this correctional to create a linear relationship that allows us to predict where a point is. it
should be created in a way that minimzes error. ANS - what is a linear regression
a residual is the value of our observed value - our predicted value(y hat) it is often what we want to
minimize, and we use the least squares regression test, that focused on outliers with higher outliers .
ANS - what is a risidual
it is basically where there are some points along our line that have a smaller or larger residual than
others, we want to try to minimize this.
also note that if there are points where all the points are above or bellow the line of best fit, it probably
indicates a non-linear relationship ANS - what is non constant error severance or heteroscedascity
questions with correct answers.
False:
establishing association is easy as long as two variables generally move together but that does not mean
they cause one another and there for have a causal link. ANS - Is the Following True or False:
Establishing causality is generally easier than establishing association
True:
the point of randomization is to make the two groups receiving treatment and caontrol as similar as
possible ANS - Is the Following True or False: Randomizing treatment and control reduces the risk of
potential confounders
False:
I am not 100% sure about this, but maybe try to think about it like smoking and cancer. smoking and
cancer are correlated, but smoking is not a sufficient cause to get it. ANS - Is the Following True or False:
if x is not a sufficient condition for y then x is not correlated with y.
False.
In the case of a list, multiplying L*2 will only result in the list being repeated twice. ANS - Is the
Following True or False: Suppose L is a list in python containing only numeric values, then L*2 will return
all the numeric values doubled.
false:
there are a lot of mistakes that can happen. Like if you are randomly selecting and you just take outliers
it will not equal the population parameter. ANS - True or False: A statistic calculated from a random
sample must be equal to the population peramiter
True:
yes that is by detention the idea of a test stat. ANS - Is the Following True or False:
The test statistic tells us the relative plausibility of the null vs alternative hypothesis.
,true ANS - Is the Following True or False
if we fail to reject the null hypothesis when the null is false we have committed a type 2 error
false:
empirical is what the distribution of what we actually observe.
sample is the distribution of what our sample predicts.
so if we roll a die 3 times that would be an empirical
if we make a computer assimilate rolling a die 3 times that would be a sampling ANS - Is the Following
True or False:
the sampling distribution is the empirical distribution of given samples values.
true:
This is what a bootstrap is ANS - Is the Following True or False:
when calculating the sample distribution for a bootstrap we use replacement with our observations
false:
95 out of 100 will be in it, but any given sample is binary so it does not have a probability. ANS - Is the
Following True or False:
a given 95% confidence interval captures the true value of a parameter with a probability of 95%
true
yeah as you add more and more data points things tend to bunch away from outliers. ANS - Is the
Following True or False:
as our sample size increases the standard dev of the sample distribution is smaller
false there are decreasing marginal returns ANS - Is the Following True or False:
The standard error decreases in a linear fashion ie increasing sample size by 2 decrease error by 2.
false ANS - Is the Following True or False:
the data used in data science is always numeric
, true
that is a feature ANS - Is the Following True or False
correlation r is unit less.
true. ANS - Is the Following True or False:
a given 90% confidence interval captures the true value of the parameter or it does not
the basic idea of association is that it is a measure of how related two variables are. it is unit less
because it just a measure of how one thing compares to another. ANS - association
in order to standardize association, we take points and subtract the mean form them and then decide
them by the standard deviation to put everything in a way that is unit less and standard and in terms of
how many standard deviations they are form the mean. also z score is centerd at 0 , and 1 is 1 stnadrd
deviationa way. this allows us to compare data ANS - how does association relate to z scores
correctional coefficient is written as r, it is a measure of how strong an association is and given between
{-1,1} association is about how two variables are related, correlation is about how two variables are
associated ANS - what is correlation, correlation coefficient and how do they relate to linearity.
it is using this correctional to create a linear relationship that allows us to predict where a point is. it
should be created in a way that minimzes error. ANS - what is a linear regression
a residual is the value of our observed value - our predicted value(y hat) it is often what we want to
minimize, and we use the least squares regression test, that focused on outliers with higher outliers .
ANS - what is a risidual
it is basically where there are some points along our line that have a smaller or larger residual than
others, we want to try to minimize this.
also note that if there are points where all the points are above or bellow the line of best fit, it probably
indicates a non-linear relationship ANS - what is non constant error severance or heteroscedascity
