Tutorial 25- Probability Density function and CDF- EDA-Data Science

[@HarshitSharma-k3p]

you are much better than many online courses in market ...thank you please keep going.

[@howtotipsandtricks4381]

Institution of Data science claims that they have good content for freshers also but it is no there, i always have to come in your channel for many topics clarification which I could never learn from there. You have a great skill of teaching😊

[@Ankurkumar14680]

Thanks for sharing sir, I always admire your teaching style, knowledge and helping nature. Small clarification, in normality plot, values on y-axis does not tell us area under the curve. In this way, y-axis corresponding to mean value on the x-axis will always be .5 but that is not the case. Actually, it is the gradient value of the CDF function (graph).

[@tensorthug6802]

The y-axis of pdf is gradient of cdf, the higher the gradients, more the density is at that particular point. The y-axis of CDF is the percentage population of a particular point.

[@MuhammadAbdullah-lr7sd]

Yes, that is what I'm thinking but in the video it creates some confusion.

[@t.saishodhanrao9519]

this comment should be pinned...... It creates lot of confusion for those who don't about this

[@jhondhelpago1638]

He explained two topics in less than 10 minutes, yet its so clear and informative

[@hadishaaben3665]

Man you do not know how much i learned from you , your explanation is AWESOME

[@im_tanmay_g]

Most simplest and non-confusing video on PDF & CDF. Thank you for the same.

[@livebiochemistry]

One of best video of PDF and CDF..thanks sir

[@mischievousmaster]

Krish loves the word PARTICULAR a ton! 😁

[@sunnyks108]

PRETTY MUCH too :)

[@srujankumar637]

you are an awesome teacher; seems like completely dwelling over the concepts... simply i take a bow

[@johan-mattias]

learning about machine learning for google spreadsheets, this helped understand the CDF so much thank!

[@cherubyGreens]

Feeling amazing with Krish Naik!

[@phalgunaa2157]

Your explanation is too good bro

[@amiysrivastava1444]

best explanation of cdf when compared with other youtube videos.

[@nvsyashwanth918]

The way you explain concepts is amazing.

[@Vidi_111]

Thank you sir .. the way you explain is so easy to understand...

[@oliullah.mahmud]

Thank you. I like your teaching style!

[@ahmedel-bahnihi346]

When you explain the PDF, you said, it is the area under the curve till that point. I think this is the CDF, not PDF. Thanks a lot for your effort nd videos

[@dharunsainath322]

that is correct...he is probably talking about CDF

[@madhuprasath6193]

A query,then how do you interpret a pdf?

[@dattamalpote2005]

i think krishna sir had explain it right.

[@mitultandon5227]

@@madhuprasath6193 It basically gives you the probability of that point. PDF would answer a question like these:- What would be the chance of weight of a person to be 90kg?. Answer to this as per the above graph in the video would be "only 25% chance ( or 0.25 probability )". Basically PDF tells us the exact probability for every point.

[@srujankumar637]

area under the curve in particular (definite integral) range is pdf:: total must be unity

[@bluestar2253]

Excellent explanation of PDF and CDF

[@PriyaAmar848]

Why do I need to use this distribution ? In which cases of data it's helpful ? Also we have uniform distribution, binomial and poisson. Where to use these. Appreciate if practical examples are included. Great explanation with graphs. Keep up the enthusias,

[@TariqueMahmud313]

Too many clear concepts in just 7 minutes !!! Thanks man!!

[@chaos8514]

hello are you also trying to learn data analysis?

[@aditisrivastava7079]

Thanks to wonderfull video..............i will simply add through pdf we can find the probababilty for a point or a range whereas cdf tell about the less than probability

[@Emotekofficial]

As far as I know CDF is Cumulative Distribution function. It can also be calculated for Probability Mass Function. But you can say in this Scenario as Cumulative Distribution function of given Probability Density Function.

[@sapnilpatel1645]

Amazing video sir. Thank you so much.

[@brayansereno4249]

Hi Krish, thank you so much, I speak Spanish but I understand you, I really need an explanation of this topic and I don't found in Spanish, you're great bro

[@navinofficial5439]

Crystal Clear!

[@kabyabasu]

Krish is the God of data science

[@bandhammanikanta1664]

Thanks for this video Krish. We will be very happy to see atleast a single reply to any comments in youtube as well as issues in his github.

[@krishnaik06]

I usually see the comments and make a note on it to create videos...github videos will be coming up soon

[@bandhammanikanta1664]

@@krishnaik06 Thank you Krishna. Waiting to see your updates.

[@himanshubansal2701]

@@krishnaik06 sir what is benefits of cdf over pdf ? bcoz we will be analysing same precentage with pdf also.

[@tsrnihar]

Hi Krish - Appreciate your effort in putting together the videos. Want to redflag something. You mixed the concepts of PDF and CDF - You are using PDF but conveying the meaning of CDF. point on a PDF indicates probability of the point in the distribution. Whereas point on a CDF indicates the cumulative probability up to the point. This is also area under the graph.

[@shrimaykher2978]

PDF doesn't indicate probability of the point in the distribution. In fact, it is very troubling to believe but probability with which continuous variable takes exact value is zero. We always talk about probability of random variable falling into an interval using PDF by finding area under curve. Moreover as sir explained, i would like to make a little correction that pdf can go beyond 1. Y-axis of PDF is probability density not % distribution. PDF is not actually probability, but rather a density function which tells amount of probability per unit length, therefore it can go beyond 1 unlike PMF. Story of discrete and continuous variable is different and we cannot mix up the theory.

[@BalaguruGupta]

I've commented on your other videos also, there I understood the love you had on the technology. From this video I understood that, actually you're so much passionate on teaching sir. The way you explained PDF and CDF is really amazing sir. Thank you so much. :)

[@VVV-wx3ui]

Simply explained. Good going Krish.

[@NR_Tutorials]

nice videos thanks krish naik sir

[@VVV-wx3ui]

Thanks Krish for sharing your knowledge. Please keep it going.

[@santamsaha9415]

this video is a soul saver

[@mukundsudharsan1294]

In 3.08 you mentioned that the y-axis in the normal distribution represents the % of distribution below that point. If that statement holds true then shouldn't the graph be continuously increasing and it would be cdf? So what does the y-axis indicate for the normal distribution falling in the right half? Please correct me if I am mistaken, but would like to understand this better.

[@kantafcb1]

y axis shows the %age distribution of intervals

[@padduchennamsetti6516]

wow you are awesome,the best

[@xeysus5907]

Great one

[@Ks-oj6tc]

Well explained, Thanks a lot Krish.

[@hbk788dbz]

PDF values can be greater than 1 ! (your y-axis can be greater than 1 ) . Area under the curve cannot be greater than 1 . Please correct me if im wrong @krish naik

[@mahalerahulm]

Excellent !! Very nice explanation.

[@AJ-et3vf]

awesome video sir! thank you!

[@muhammadyasirbutt3631]

very well brother your are great teachning

[@jaheerkalanthar816]

Thanks for the video sir, I learned lot of things in this video

[@karthikvijayasarathi89]

Small correction - Probability cannot go over 1 , but probability density function can go. Correct me if I am wrong

[@aimenbaig6201]

you are the best

[@magedrefat1658]

Sir, your explanation is amazing ^_^

[@ArchnaVijay]

amazing video

[@kpratik5551000]

Very good explanation.

[@alexv1602]

Your god for me now😮

[@vidyasurbhi3084]

Very nice 👌.

[@saurabhtripathi62]

thanks your series is great , u made this very easy.

[@mohammedabdulahmed8808]

Simply amazing explanation 😍😍 Thanks alot and keep doing sir!!!

[@rakhijha8911]

Simply amazing i read so many articles on cdf but everyone was calculating the value no one explained it so well can I plzz connect you on LinkedIn

[@DataAI_junction]

thank you for the video

[@TheAl217]

Thank you for clarifying these functions.

[@ranjanpal7217]

Amazing...plz make a video on how to determine the distribution of a dataset using Python.

[@sagarkhande4412]

Ty sir your video was really helpful..👍

[@equbalmustafa]

Nice one

[@shadiyapp5552]

Thank you sir ♥️

[@ankitchakraborty1126]

Hi sir. Thanks for sharing awsome content like this. I have 1 question. Can we calculate percentile and median from CDF?

[@nehasaroha2505]

Very informative video....can you suggest some simpler kaggle datasets on which we can perform EDA using PDF,CDF and multivariate analysis. I have already done on iris, Titanic and Haberman's dataset, but was thinking about getting more practice.

[@muhammadyasirbutt3631]

hello pretty friend

[@muhammadsaqib2961]

Good explanation

[@গোলামমোস্তফা-শ৮থ]

But that area is not 1 according to your curve. Because if the interval is 10 and and suppose height of any bar is 0.25 then 10*0.25 > 1. May be The concept exactly came from CDF function and in pdf the values are called likelihood and these values are the slopes(tan@) of different points from cdf.

[@deepakbhaiya.shorts]

Old is gold

[@tsaurav18]

thankyou sir.

[@SandeepGurjar-ko5ju]

Hi, nice video but I think while you were explaining PDF you were actually talking about CDF.

[@ahmed96616]

Excellent !

[@saddamshaikh9285]

Sir please make a video on navie byes algorithm.

[@prachetade3163]

Hi Krish, I think you meant to say that cdf is the value of integration of the pdf uptil a point x and not a summation of pdf as you explained at 4.47. Great videos btw, really appreciate your efforts.

[@amolkabugade3728]

integration is actually addition of those values

[@KalyanGk0]

Great explanation krish😊 .please make video on practical implementation of this concepts using python.

[@schuf1738]

Thank you !!

[@lokeshchoraria6559]

best explanation

[@sudhirBhalekar007attaboy]

Why do we calculate CDF as PDF is already giving you % of distribution for required data analysis, through this CDF, we are getting added (C.values) but what is the significance of this concept?

[@rahulmandalsky]

When you are Finding the %density at any point X, what you are explaining is the Cumulative density, not probability density, from a initial point to point Xi.

[@umasharma6119]

Probability density at that specific point.

[@ibrahimahmethan586]

thank u so much . god bless u

[@MScFabianoBriao]

Buenos! Do you have videos of real cases showing inferential statistics to test (validate) models?

[@menakask6050]

Hi krish, kindly let me know pls explain how do you say at point 130 in X axis with 90% in distribution in Y axis is "less than" since the CDF is straight it is increasing and you are mentioning less than 130kg is there in 90% of the dataset. How do you predict it is less or high using CDF?

[@sharathchandrakarnati4615]

We can use logistics regression right based upon cdf ?

[@pushkarsaini7653]

sir may u please make video on hypothesis.

[@himanshusethi8246]

So basically what CDF does is that it adds up all the previous % values(in the PDF) from before the given x-coordinate value and gives us the plot of those values. Please do correct me if I am wrong

[@PriyaAmar848]

yes - cumulative itself says add all values till the point targeted . Hence curve raises beyond PDF

[@naveenreddy7421]

Sir it is clear but please make a video on same concept in jupyter note book

[@nightowl1596]

elite explanation, gg

[@Nikhil-jj7xf]

Krish pls provide you're online course registration link

[@saurabhtripathi62]

please add a video that clears how to do this with python , practically

[@shreyasaxena5169]

sns.distplot(df['weight])

[@0505Arjun]

What scenarios we will use CDF and PDF in machine learning..?

[@gauravsaini728]

I have the same query. Can you please give any domain specific example and show is how PDF and CDF curves will help the data scientist to take certain decisions..

[@alphonseinbaraj7602]

yeah ..same query me too

[@rajatchaturvedi7379]

Don't know its application in ML yet, Since I have started learning recently but one use of it is in EDA Which is the exploratory data analysis.Exploratory means you don't know anything about the data set from before. Your task is to extract some basic yet critical information about the dataset before implementing ML algorithms.Its important to get an idea about the dataset .PDF, CDF various plots such as 2D, pair plot , etc are some aspects of EDA.There are other stuff also.You can google it to understand more .

[@amitmaurya6179]

If you are smoothening the curve, why count changes to percentage of distribution.

[@rakeshenjapuri3143]

how will calculate the above 60% in pdf and how will take the cumulative percentage in cdf give with mathematical explanation sir

[@jaychauhan2933]

Can you please show the practical implement of PDF and CDF

[@wealth_developer_researcher]

Amazing :)

[@dude5697]

In 3.05, you shade the area even outside of the bell curve, but in 3.31 you shade the area only within the bell curve. I can't understand , Pls can you explain me Krish sir?

[@roshankumargupta9978]

Thanks for the video Krish. But I do have a doubt that in a long run.. how could be it will be beneficial that this much percentage of data lies upto certain threshold?

[@yaminimadan4087]

outlier detection perhaps

[@lahari1512]

hai krish , ur vedios are really helping me to learn machine learning very easily , can u please upload svm and xg boost vedios please its a request

[@Thanusree234]

Sir is this subject exploratory data analysis and statistics subject the same sir please reply 🙏

[@yashikaarora8573]

what's actually probability density function sir? this question I don't think is yet solved CDF is clear but not pdf watched a lot video but did not understand its reference with data science

[@adityapathania3618]

dont you think the cumulative total will go above 1? at 3rd 4th value itself ?as the probabilities are getting added ?

[@ujjwalmv9697]

what if cdf over a period of time is not being constant and hits 90 degree and then goes constant, why is that straightline coming in cdf?

[@anto1756]

Nice 😁 could u do a comparison about survival function, inf and all other methods please

[@umesh789s]

can you please explain about T-score and T distribution

[@baijuthomas3716]

so is CDF simply a another representation of a PDF , just wondering at what point do you decide to use a CDF over a PDF if it simply represents data in a different way ? it seems like the CDF always locks the results between 1-0 . what if the returns go -Negative how would you represent that ? Thanks as always .

[@bavalpreetsingh4664]

do make video on all the distributions