Reputation
11,976
Top tag
Next privilege 15,000 Rep.
Protect questions
Badges
1 14 37
Newest
 self-study
Impact
~492k people reached

Feb
5
comment Compute accuracy of model evaluation
TP is "True Positive" (predicted positive and actually positive: $0$ here). TN is "True Negative" (predicted negative and actually negative: $9990$ here). FP is "False Positive" (predicted positive and actually negative: $0$ here). FN is "False Negative" (predicted negative and actually positive: $10$ here).
Feb
5
comment Why is the probability of my chi square statistic equal to 0
Apparently not R. For example 1-pchisq(90,8,lower.tail=TRUE) gives 4.440892e-16 though this has started to have precision issues and it would be better to use pchisq(90,8,lower.tail=FALSE) giving 4.650448e-16
Feb
3
comment Compute the % using SD, and Mean
You will probably need a <= or something similar in your calculation
Jan
28
comment Why is the logistic distribution called “logistic”?
For me, the association is with the log-odds or logit function $\log\left(\dfrac{p}{1-p}\right)$ which has the inverse $\dfrac{\exp(x)}{1+\exp(x)} = \dfrac{1}{\exp(-x)+1}$ which is the standard logistic function. So is it log- as in "logarithm" and -istic (or -istique in French) as in "related to".
Jan
25
comment Do I need a special kind of linear regression for aggregated data?
Another issue is that some states are bigger than others
Jan
25
comment simulate data for different sample size with same parameter
@NorHishamHaron It is unreasonable to expect a sample to have the same skewness and kurtosis as the population. I have added a graph illustrating this
Jan
23
comment How to check if a distribution has undefined variance?
Related: stats.stackexchange.com/questions/2504/test-for-finite-variance
Jan
23
comment Rate at which a Gaussian random variable is the maximum in a set of independent Gaussian random variables
If they were identically distributed (Gaussian with same mean and variance), as well as independent, it would be $\frac1n$
Jan
19
comment How does boxplot in R calculate quantiles?
You might want to add fivenum(a); abline(v=fivenum(a), col="green") to the end of your code and see the effect
Jan
13
comment Probability theory
You do not need to consider the sigma-algebra explicitly: you can just treat the status of each integer $i$'s membership of $A$ and $B$ to be independent of the other integers' statuses, so take the answers when $n=1$ and raise them to the power of $n$
Jan
7
comment Is it a good practice to always scale/normalize data for machine learning?
Related: stats.stackexchange.com/questions/89809/…
Jan
1
comment Why does Random Forest use randomness at all?
If you have a large enough set of data, you will hit your "up to some limit" rather quickly
Dec
9
comment Probability function of two coin flips, is $P=\frac{1}{4}$ valid?
If $X$ is a function $\Omega \to S$ then presumably so.
Dec
9
comment Probability function of two coin flips, is $P=\frac{1}{4}$ valid?
If you are saying $\displaystyle \sum_{s \in S} P_X^{}(s) = 1$ then this is correct and corresponds to your earlier $\displaystyle\sum_{w\in\Omega}P(w)=1$
Dec
9
comment Intuitive explanation for variance of vector
The dot product of $X$ and $Y$ is not $1$ "when the vectors are parallel", as it depends on the length of the vectors: you may be mixing covariance and correlation.
Nov
26
comment Probability Questions within R
Personally I use the default lower=TRUE and subtract from 1 if necessary, as this corresponds to the CDF. Your answer to (iii) could either use sum as in (vi) or use the difference of two pbinom. Your answer to (x) should be similar I suspect is currently wrong. I have not checked the others
Nov
26
comment Check if distribution is bell-shaped
What so you have? A density function? A sample?
Nov
11
comment How can I recreate a Weibull distribution given mean and standard deviation and the shape and scale parameters are unknown?
@soakley: It may depend on your version of Excel. MIne (Excel 2013) does have a direct Gamma function: =GAMMA(5) gives 24
Nov
10
comment t-distribution having heavier tail than normal distribution
A graph with $\log S(x)$ (and perhaps extending $x$ a little) might demonstrate the heavier tails more clearly, and could also work with higher degrees of freedom,
Nov
10
comment t-distribution having heavier tail than normal distribution
Though for a $t$-distribution with $3$ or $4$ degrees of freedom, the kurtosis is infinite, while with $2$ degrees of freedom the standard deviation is infinite so you cannot calculate the kurtosis, and with $1$ degree of freedom you cannot even calculate the mean or the $4$th moment.