00:08 Covering basic and advanced topics related to data science positions 02:32 Understanding Statistical Analysis 07:00 Descriptive stats is the organizing and summarizing of data. 09:17 Understanding inferential statistics and the difference between population and sample. 13:37 Different sampling techniques and their importance 15:56 Stratified sampling is a technique where the population is split into non-overlapping groups. 19:49 There are two sampling techniques: random sampling and convenient sampling. 21:41 Sampling techniques may vary based on the use case. 25:24 There are two types of variables: quantitative and qualitative. 27:25 Classification of variables based on characteristics, such as IQ and t-shirt size 31:21 There are four types of measurement wells that include nominal, ordinal, interval, and ratio related data. 33:15 Ordinal, Interval, and Ratio data types explained 36:55 Bar graphs and pie charts can be used to represent discrete variables. 38:52 Histograms are used to represent continuous values through bins. 42:48 Arithmetic mean is the average of a specific distribution. 44:43 Mean, Median, and Mode are the three main measures of central tendency. 48:30 Outliers have a major impact on data distribution 50:22 The median is a measure of central tendency that is not affected by outliers 54:19 Mode is used to handle missing values and find the most frequent element. 56:19 Suitable measure for ages of students 1:00:18 The calculated value is 1.81 1:02:20 Variance measures the spread of data. 1:06:24 Standard deviation and variance are important in understanding data spread. 1:08:18 Finding outliers and understanding percentiles 1:12:34 80% of the distribution is less than 10 1:14:35 The five number summary is used to analyze data and remove outliers. 1:18:35 Compute the lower fence and the higher fence values. 1:20:41 The 5 number summary for the given data is: 1, 3, phi, 7, 9. 1:24:50 Summary of Statistical Distributions 1:26:55 Distributions are a way to visualize continuous data. 1:30:50 The normal distribution is important for deriving conclusions. 1:32:47 Empirical formula helps in understanding the distribution of data 1:36:45 4.75 falls 0.75 standard deviation to the right of the mean 1:38:47 The z-score helps calculate standard deviations and their positioning on a bell curve. 1:42:51 Convert data into standard normal distribution using z-score 1:44:58 Standardization and normalization are two different processes used for data conversion. 1:49:03 The average score of Rishabh Pant in the series was 260. 1:51:02 The average score of the series is -1.25 1:55:29 The standard deviation of the scores indicates the distribution pattern. 1:57:29 The main question is the percentage of scores that fall above 4.25. 2:01:19 Z table shows area to the right of the curve 2:03:27 The left and right areas can be calculated by subtracting from the mean and standard deviation. 2:07:26 Compute the z score and find the area under the curve. 2:09:21 Understanding body area symmetry and how to compute mean 2:13:18 The data distribution does not follow a Gaussian distribution. 2:15:40 Discussing topics on IQR, probability, permutation and combination, confidence intervals, p-value, and hypothesis testing. 2:19:30 Implementing the outliers detection function using the z-score formula in Python. 2:21:26 Finding outliers using z-score formula 2:25:19 Outliers can be identified using z-score and interquartile range (IQR). 2:27:18 Find the lower and upper fence using q1 and q3 respectively. 2:31:17 Box plot creation and importance of probability in machine learning 2:33:05 Probability can be defined as the number of ways an event can occur divided by the number of possible outcomes. 2:36:48 Multiple events can occur at the same time, such as drawing a king or a queen from a deck of cards. 2:38:42 Probability of mutually exclusive and non-mutually exclusive events 2:42:38 Understanding probability concepts: addition rule and multiplication rule 2:44:35 Events are independent and do not impact each other 2:48:22 Probability of drawing a queen and an asus from a deck of cards 2:50:18 Conditional probability helps in biased theorem 2:54:09 Permutation and Combination in Mathematics 2:56:04 P-Value 3:00:05 The coin is fair. 3:02:07 Hypothesis testing involves four steps: proof, fairness of coin, alternative hypothesis, and experiment 3:06:17 The significance value of 0.05 is used to determine if a coin is fair or not. 3:08:18 The coin is not fair. 3:12:04 Type 1 and Type 2 errors in hypothesis testing 3:13:58 Rejecting the null hypothesis can be a good or bad decision depending on whether the alternate hypothesis is true or false 3:17:52 Outcome four is accepting the null hypothesis when it is true. 3:19:44 One-tailed and two-tailed test explained. 3:23:38 The experiment is conducting a two-tailed test on the placement rate of a college. 3:25:30 Confidence interval is important in statistical analysis 3:29:14 Confidence intervals help determine the range around the population mean 3:31:17 Construct a 95% confidence interval about the mean 520. 3:35:16 The upper bound of the confidence interval is 12947.52 3:37:18 The confidence interval for the average size of sharks throughout the world is 520 with a lower bound of 480.8 and an upper bound of 559.2. 3:41:24 Population standard deviation is not given, so we use t test. 3:43:33 Compute the lower bound and upper bound using the sample variance problem and t-table. 3:47:45 Researchers want to test a new medication to see its effect on intelligence. 3:49:45 The hypothesis test is a two-tailed test with a confidence interval of 95%. 3:53:38 Standard error is calculated by dividing the standard deviation by the square root of the sample size. 3:55:31 The mean is not equal to 100. 3:59:44 T-test is used to compare means of two groups 4:01:51 The t value is greater than the decision rule, indicating the rejection of the null hypothesis. 4:06:22 Chi square test is a non-parametric test performed on categorical or ordinal data. 4:08:30 In 2010, the distribution of ages in a small town has changed compared to 2000 census. 4:12:34 The observed distribution of the population is less than 18: 20%, 18 to 35: 30%, and greater than 35: 50%. 4:14:40 There is a huge difference between the observed data and the expected distribution based on the 500 samples. 4:18:37 The chi square test statistic is 232.94, which is greater than 5.99. 4:20:48 Performing z test using Python to determine the significance of a new drug on IQ level 4:24:34 Covariance and significance value 4:26:37 The significance level determines whether to accept or reject the null hypothesis. 4:30:29 Positive correlation between x and y when x is increasing y is also increasing, negative correlation when x is decreasing y is also increasing, no relationship between x and y when covariance is 0 4:32:17 Covariance and Pearson correlation restrict values between -1 and +1 4:36:00 Covariance and correlation capture the linear properties, but Spearman rank correlation also captures non-linear properties. 4:37:51 The formula for calculating the Spearman rank correlation 4:41:48 Performing a one-sample t-test with a sample size of 10 to determine if the mean is close to the population mean. 4:43:52 The example demonstrates the changes in values based on different scenarios. 4:47:42 The results should not be rejected as the p-value is extremely low. 4:49:50 Reject the null hypothesis if p value is less than or equal to 0.05. 4:53:55 Discussing various distributions and their significance 4:55:57 The mean weight of 36 individuals is 169.5 pounds. 5:00:03 The area under the curve is 0.99 triple 1 5:02:02 The calculated p-value is 0.0089. 5:06:11 The z-score is 2.30, indicating rejection of the null hypothesis. 5:08:28 The average age of a college is 24 years with a standard deviation of 1.5. 5:12:29 The p-value is significantly smaller than the significance value, indicating rejection of the null hypothesis. 5:14:35 Bernoulli distribution is a probability distribution with two outcomes: 0 or 1. 5:18:10 Probability Mass Function (PMF) explained for categorical variables 5:19:57 Binomial and Pareto distributions are important in statistics. 5:23:49 Log-normal distribution and its relationship with power law distribution 5:25:39 The distribution follows a Pareto distribution and can be converted to a normal distribution using the central limit theorem.
Best statistics course ever. I was looking for statistics courses for ML that explained concepts in a way that didn't drag on, and gave examples easy for high schoolers to understand. Thank you sir!
Absolutely Magnificent !! Just gone through the full video and I must say it's very informative. All the best I hope you get more success and happiness! Thank you so much!
Really, your explanation is too good. When I read the topics, I understood them in one way, but after watching your videos, I could think in a more practical way and see when we could apply them.
Till now I have completed 3 hrs of this video & it is extremely helpful. Krish Sir, I must say that you know the art of explaining complex thing in simplest way. Thank for making this kind of helpful videos.
Exactly this, he got it very confused in the video. For other folks who are confused: at 95% confidence, alpha = 0.05, at 90% confidence, alpha = 0.1, this is within the confidence interval if alpha = 0.05 Looking at this, if alpha is 0.05, then a value > 0.05 (cuz 0.1 > 0.05) will fall within the confidence interval. So, p>alpha implies that it lies within the confidence interval, so we accept the null hypotheses. p
Sir i cant thankyou enough , I greatly appreciate your way of teaching. My interview is close and i pretty much covered everything i wanted from your channel alone. Thankyou Krish
thankyou so much sir for this wonderful course statistics video...i came here as a newbie but now after completing this i have much knowledge about data and also my approach has changed in seeing data and treating it...blessed to find your channel thankyou sir❤❤
@connecttechrockstar4474 😂😅i just created this channel to comment on videos. While creating the account I couldn't think of any other name, and so I kept this. 😅
I bought the Data Science and Machine Learning course from GFG, in the statistics section the faculty suddenly started speaking alien language, and now I am here enjoying the best free content.
You are a Genius and I love the way you teach. Most of the topiccs I never really understood during my 5 years in school, it became so easy and simple for me. Thank you so much. You are indeed a blessing.
There are multiple types of distribution we have to learn some of them are bionimial distribution, gaussian distribution, Geometric distribution , exponential distribution , gamma distribution , beta distribution, Poisson distribution,weibull distribution,cauchy distribution
very informative and helpful video for revising all concept and also for interview purpose thank you krish sir for making for us this like beautiful video
3:58:00 : if sample mean = 110, then Z = 3.65 & it's not in our calculated UB & LB range (-1.96 to +1.96) So, we reject the Null Hypothesis & it improves the Intelligence as Z > UB.
Sir, I am going to do bstat in isi but I had not statistics at 10+ 2 level. Your video has helped me to start my journey with statistics by making every concept crystal like clear.
Hello, for those who are looking for notes, you can go to one of his original live videos and from the description box you can go to the course dashboard and from there you can get the notes for each day separately in the resources section.
1:24:00 = Why sample variance is divided by n-1? ua-cam.com/video/vGsRwB3TsiE/v-deo.html Summary : Researchers found that using the denominator (n-1) in sample variance/standard deviation calculations provides estimates closer to the population variance/standard deviation in various types of sample data distributions (positively/negatively skewed). This correction is also known as Bessel's correction or degrees of freedom.
sir with due respect you have made a major mistake in P-value and significance value Hypothesis conclusion 4:48:39 - A p-value less than 0.05 (typically ≤ 0.05) is statistically significant. It indicates strong evidence against the null hypothesis, as there is less than a 5% probability the null is correct (and the results are random). Therefore, we reject the null hypothesis, and accept the alternative hypothesis. You did the opposite. Overall the course was good but minor mistakes here and there. Thank you Though you corrected it later.. But best is when editing you could just add * and add comment on video. Because while studying and taking notes with lecture it becomes a very bad experience to go back and correct all the wrong. The thought process needs to changed fully to understand it again..
@@bapupatil9354hi did you have any prior experience, if not kindly share how did you apply for the job because I am fresher & trying to apply for job but not getting any calls as well as emails kindly support
On the measures of central tendency, I noticed, you kept referring to the median as the mode. Those are two different measures. Luckily you corrected it down the line. But overall, perfect summary of statistics.
First of all thanks for producing such a useful and insightful video on Statistics. Now my question is about exit poll results (almost all are failing). What I infer from that is: 1) Samples are biased- As they claim of random sampling but the samples are biased (gender biased, community biased, wrong answering biased). 2) Sample size- owing to humongous population of our country its quite impossible to collect even considerable sample data from all types of populations. 3) Biasing in result predictions- as can be seen all analysis of exit poll is agenda driven that is party specific. (Human bias) 4) Collection techniques- as technology progresses still these companies rely on old conventional way of inferring the results. Most of them rely on structured data or on survey reports but in todays Data driven world unstructured data can predict better results which all companies are lagging.
2:08:07 , we can consider this approach , where the area between range 100 till 145 (right part) to be 50 % as the guassian distribution is generally a symmetric one. and we also know , area between 85 to 115 is 68 % as it the between -1 SD to +1SD as per 68% 95% 99.7% rule , there fore area between 85 to 100 would 34 (68/2) due to its symmetry. now on adding area between (85-100 and 100-115), we get 50+34 = 84 . 84% represents the region of people having IQ more than 85. to find less than 85 we need to subract 100 from it, i.e. 100-84 = 16%
02:24:31 In my case it will show empty list of outliers because 100, 110, and 115 are not extreme enough to be classified as outliers with a threshold of 3. So i change the threshold 3 to 2 this way it will work. My dataset is : dataset = [50, 52, 48, 47, 51, 49, 53, 45, 46, 54, 55, 50, 49, 52, 47, 48, 50, 51, 46, 53,100,110,115]
Watched the video in a day and a half, to revise stats that I had last studied in engineering days, helped me a lot cracking 3 quant interviews ending up in a 50% higher pay quant researcher role. Thanks brother. ❤
Hi sir recently i ve seen this video !! So amazing and helpful for my data science carrer !! Now i saw in other videos that in hypothesis testing , the type 1 error is false positive and type 2 error is false negative
Previously i had sawn a lot of vedios sir no vedio given me the proper understanding .Thanks a lot sir not only understandig,the vedio also given me with the examples of the interview questions thanks thanks a lot sir🥰🥰🥰
Hello Sir, in chi square census example ,why we need to choose the 0.05 in the chi square table ? actually it is two tail test right! so we need to see the value of 0.025 in the table with respect to df = 2 right ?
4:48:49 if p-values is less than alpha, then we need to reject the null hypothesis. This needs to be corrected in the video. Let me know if I missed something. It was a great explanation overall. Thanks Later in the video, correction has been made. Thanks
thank you for the video. I have a subject named Introduction to Data Science using python and has statistic in my syllabus, as it was only introduction, i have completed the video till t - test and i have basic idea about covarience and correlation. My exam is tomorrow let's see if i pass.
Exactly this, he got it very confused in the video. For other folks who are confused: at 95% confidence, alpha = 0.05, at 90% confidence, alpha = 0.1, this is within the confidence interval if alpha = 0.05 Looking at this, if alpha is 0.05, then a value > 0.05 (cuz 0.1 > 0.05) will fall within the confidence interval. So, p>alpha implies that it lies within the confidence interval, so we accept the null hypotheses. p REJECT THE NULL HYPOTHESIS p > alpha (domain expert will tell this values) ---> ACCEPT THE NULL HYPOTHESIS
5hr:3min, Each tail will have 0.00889 and the middle region will have 0.9822, Z value gives the region below 2.307 which also includes area less than -2.307. So we need to subtract that tail value to get the middle region.
I agree with with u I also think the same way bcoz while finding out the z score we r getting the area from +2.307 till -ve end not till -2.307, so i guess it should be each tail as 0.00889 which means P-value is 0.00889*2 and the middle region will be 1-(0.00889*2). But with all due respect, hats-off to u Krish sir for making stats soo simple to under. Because of u only we are able to spot even these minor things
Hi Krish Sir, Thanks for the amazing informative video. In type 1 and type 2 error while explaining confusion matrix TN should be type 2 error. Earlier you wrote correctly, but later you marked FN as type 2 error which is incorrect.
dear sir, with due respect i want to ask something in case when we use trim mean and normal mean to find outlier and triimmean value is very close to normal mean ,it is said that outlier is not present in that particular table but when we check with quartile method there exist a outliers in the given table
00:08 Covering basic and advanced topics related to data science positions
02:32 Understanding Statistical Analysis
07:00 Descriptive stats is the organizing and summarizing of data.
09:17 Understanding inferential statistics and the difference between population and sample.
13:37 Different sampling techniques and their importance
15:56 Stratified sampling is a technique where the population is split into non-overlapping groups.
19:49 There are two sampling techniques: random sampling and convenient sampling.
21:41 Sampling techniques may vary based on the use case.
25:24 There are two types of variables: quantitative and qualitative.
27:25 Classification of variables based on characteristics, such as IQ and t-shirt size
31:21 There are four types of measurement wells that include nominal, ordinal, interval, and ratio related data.
33:15 Ordinal, Interval, and Ratio data types explained
36:55 Bar graphs and pie charts can be used to represent discrete variables.
38:52 Histograms are used to represent continuous values through bins.
42:48 Arithmetic mean is the average of a specific distribution.
44:43 Mean, Median, and Mode are the three main measures of central tendency.
48:30 Outliers have a major impact on data distribution
50:22 The median is a measure of central tendency that is not affected by outliers
54:19 Mode is used to handle missing values and find the most frequent element.
56:19 Suitable measure for ages of students
1:00:18 The calculated value is 1.81
1:02:20 Variance measures the spread of data.
1:06:24 Standard deviation and variance are important in understanding data spread.
1:08:18 Finding outliers and understanding percentiles
1:12:34 80% of the distribution is less than 10
1:14:35 The five number summary is used to analyze data and remove outliers.
1:18:35 Compute the lower fence and the higher fence values.
1:20:41 The 5 number summary for the given data is: 1, 3, phi, 7, 9.
1:24:50 Summary of Statistical Distributions
1:26:55 Distributions are a way to visualize continuous data.
1:30:50 The normal distribution is important for deriving conclusions.
1:32:47 Empirical formula helps in understanding the distribution of data
1:36:45 4.75 falls 0.75 standard deviation to the right of the mean
1:38:47 The z-score helps calculate standard deviations and their positioning on a bell curve.
1:42:51 Convert data into standard normal distribution using z-score
1:44:58 Standardization and normalization are two different processes used for data conversion.
1:49:03 The average score of Rishabh Pant in the series was 260.
1:51:02 The average score of the series is -1.25
1:55:29 The standard deviation of the scores indicates the distribution pattern.
1:57:29 The main question is the percentage of scores that fall above 4.25.
2:01:19 Z table shows area to the right of the curve
2:03:27 The left and right areas can be calculated by subtracting from the mean and standard deviation.
2:07:26 Compute the z score and find the area under the curve.
2:09:21 Understanding body area symmetry and how to compute mean
2:13:18 The data distribution does not follow a Gaussian distribution.
2:15:40 Discussing topics on IQR, probability, permutation and combination, confidence intervals, p-value, and hypothesis testing.
2:19:30 Implementing the outliers detection function using the z-score formula in Python.
2:21:26 Finding outliers using z-score formula
2:25:19 Outliers can be identified using z-score and interquartile range (IQR).
2:27:18 Find the lower and upper fence using q1 and q3 respectively.
2:31:17 Box plot creation and importance of probability in machine learning
2:33:05 Probability can be defined as the number of ways an event can occur divided by the number of possible outcomes.
2:36:48 Multiple events can occur at the same time, such as drawing a king or a queen from a deck of cards.
2:38:42 Probability of mutually exclusive and non-mutually exclusive events
2:42:38 Understanding probability concepts: addition rule and multiplication rule
2:44:35 Events are independent and do not impact each other
2:48:22 Probability of drawing a queen and an asus from a deck of cards
2:50:18 Conditional probability helps in biased theorem
2:54:09 Permutation and Combination in Mathematics
2:56:04 P-Value
3:00:05 The coin is fair.
3:02:07 Hypothesis testing involves four steps: proof, fairness of coin, alternative hypothesis, and experiment
3:06:17 The significance value of 0.05 is used to determine if a coin is fair or not.
3:08:18 The coin is not fair.
3:12:04 Type 1 and Type 2 errors in hypothesis testing
3:13:58 Rejecting the null hypothesis can be a good or bad decision depending on whether the alternate hypothesis is true or false
3:17:52 Outcome four is accepting the null hypothesis when it is true.
3:19:44 One-tailed and two-tailed test explained.
3:23:38 The experiment is conducting a two-tailed test on the placement rate of a college.
3:25:30 Confidence interval is important in statistical analysis
3:29:14 Confidence intervals help determine the range around the population mean
3:31:17 Construct a 95% confidence interval about the mean 520.
3:35:16 The upper bound of the confidence interval is 12947.52
3:37:18 The confidence interval for the average size of sharks throughout the world is 520 with a lower bound of 480.8 and an upper bound of 559.2.
3:41:24 Population standard deviation is not given, so we use t test.
3:43:33 Compute the lower bound and upper bound using the sample variance problem and t-table.
3:47:45 Researchers want to test a new medication to see its effect on intelligence.
3:49:45 The hypothesis test is a two-tailed test with a confidence interval of 95%.
3:53:38 Standard error is calculated by dividing the standard deviation by the square root of the sample size.
3:55:31 The mean is not equal to 100.
3:59:44 T-test is used to compare means of two groups
4:01:51 The t value is greater than the decision rule, indicating the rejection of the null hypothesis.
4:06:22 Chi square test is a non-parametric test performed on categorical or ordinal data.
4:08:30 In 2010, the distribution of ages in a small town has changed compared to 2000 census.
4:12:34 The observed distribution of the population is less than 18: 20%, 18 to 35: 30%, and greater than 35: 50%.
4:14:40 There is a huge difference between the observed data and the expected distribution based on the 500 samples.
4:18:37 The chi square test statistic is 232.94, which is greater than 5.99.
4:20:48 Performing z test using Python to determine the significance of a new drug on IQ level
4:24:34 Covariance and significance value
4:26:37 The significance level determines whether to accept or reject the null hypothesis.
4:30:29 Positive correlation between x and y when x is increasing y is also increasing, negative correlation when x is decreasing y is also increasing, no relationship between x and y when covariance is 0
4:32:17 Covariance and Pearson correlation restrict values between -1 and +1
4:36:00 Covariance and correlation capture the linear properties, but Spearman rank correlation also captures non-linear properties.
4:37:51 The formula for calculating the Spearman rank correlation
4:41:48 Performing a one-sample t-test with a sample size of 10 to determine if the mean is close to the population mean.
4:43:52 The example demonstrates the changes in values based on different scenarios.
4:47:42 The results should not be rejected as the p-value is extremely low.
4:49:50 Reject the null hypothesis if p value is less than or equal to 0.05.
4:53:55 Discussing various distributions and their significance
4:55:57 The mean weight of 36 individuals is 169.5 pounds.
5:00:03 The area under the curve is 0.99 triple 1
5:02:02 The calculated p-value is 0.0089.
5:06:11 The z-score is 2.30, indicating rejection of the null hypothesis.
5:08:28 The average age of a college is 24 years with a standard deviation of 1.5.
5:12:29 The p-value is significantly smaller than the significance value, indicating rejection of the null hypothesis.
5:14:35 Bernoulli distribution is a probability distribution with two outcomes: 0 or 1.
5:18:10 Probability Mass Function (PMF) explained for categorical variables
5:19:57 Binomial and Pareto distributions are important in statistics.
5:23:49 Log-normal distribution and its relationship with power law distribution
5:25:39 The distribution follows a Pareto distribution and can be converted to a normal distribution using the central limit theorem.
Sir not only took pain to remove unnecessary parts, he also sped up video a little to save students' time...Hats off
Is this good or bad?
@@lazydamsel obviously good
@@ShauriePvs cool cool. I'll watch. Thanks!
how can I get the notes. It states unable to render code on Github.@@ShauriePvs
Best statistics course ever. I was looking for statistics courses for ML that explained concepts in a way that didn't drag on, and gave examples easy for high schoolers to understand. Thank you sir!
Watching Your Videos is better than Watching Netflix. Thank you sir.
Does Netflix have commercials every 5 seconds
@@loujon191Do you pay to watch yt?
@@loujon191 Does Netflix provide it's content for free ??
@@loujon191You pay for netflix, if you pay for youtube premium tou won’t have them
@@loujon191but you have to pay for subscription mate !
Thank you for uploading all basic statistics in one video
You really explained all concepts in a single video
Thank you for making stat so easy to understand. Awesome all in one content.
Absolutely Magnificent !! Just gone through the full video and I must say it's very informative. All the best I hope you get more success and happiness! Thank you so much!
Really, your explanation is too good. When I read the topics, I understood them in one way, but after watching your videos, I could think in a more practical way and see when we could apply them.
Thank you for putting this all in one summary. What you do here and ineuron is off the charts (I am a proud subscriber)
Bhai apni video ky metereils download keye hai agar to muja b sand kardo plz mai download karni ky koshish ke lekin download nahi hoty
no words to say. Legendary content!!!
Till now I have completed 3 hrs of this video & it is extremely helpful.
Krish Sir, I must say that you know the art of explaining complex thing in simplest way. Thank for making this kind of helpful videos.
4:26:50 : Key point to note, if P-value
No,
If the p-value is greater than alpha, you accept the null hypothesis. If it is less than alpha, you reject the null hypothesis.
Exactly this, he got it very confused in the video. For other folks who are confused:
at 95% confidence, alpha = 0.05,
at 90% confidence, alpha = 0.1, this is within the confidence interval if alpha = 0.05
Looking at this, if alpha is 0.05, then a value > 0.05 (cuz 0.1 > 0.05) will fall within the confidence interval.
So, p>alpha implies that it lies within the confidence interval, so we accept the null hypotheses.
p
@@sajankumarkar8237 yeah you are right he got it confused
he corrects it later in the video at 4:49:53
Sir i cant thankyou enough , I greatly appreciate your way of teaching. My interview is close and i pretty much covered everything i wanted from your channel alone. Thankyou Krish
Thank you sir for this awesome tutorial and your teaching skills is just amazing 😍
thankyou so much sir for this wonderful course statistics video...i came here as a newbie but now after completing this i have much knowledge about data and also my approach has changed in seeing data and treating it...blessed to find your channel thankyou sir❤❤
Super sir. What I never understood @ school for 4 years. You have taught me in 6 hours. You are amazing. Thank you.
My Statistics Revision: Completed video in a day, amazing content, everything covered! Thanks, Krish sir and team.
is this amount of statistics enough for a entry level data scientist?
Yeah, In my opinion.
@@bapupatil9354 okay
@connecttechrockstar4474 😂😅i just created this channel to comment on videos. While creating the account I couldn't think of any other name, and so I kept this. 😅
I bought the Data Science and Machine Learning course from GFG, in the statistics section the faculty suddenly started speaking alien language, and now I am here enjoying the best free content.
Please tell whether that course is worth to do or not or simply can we update our skills??!!
@@user-du5ed7of6j stay away from gfg data science and machine learning course
Thanks Krish Sir, your videos cleared my concepts too much.... 🥰🥰🥰🥰
Hey Man. I've been watching many of your videos and they are super helpful. Thank you sincerely.
You are a Genius and I love the way you teach.
Most of the topiccs I never really understood during my 5 years in school, it became so easy and simple for me.
Thank you so much. You are indeed a blessing.
Waoh! this has been my best tutorial on Statistic so far... Thank you so much for your explanation
Thank you Krish!! Learned a lot from this session.
There are multiple types of distribution we have to learn some of them are bionimial distribution, gaussian distribution, Geometric distribution , exponential distribution , gamma distribution , beta distribution, Poisson distribution,weibull distribution,cauchy distribution
Thank you sir for this awesome tutorial and your teaching skills is just amazing
Hi Krish can you please provide the materials used in this course i tried to access it using the link in the description but i'm receiving an error.
You are the best teacher i have in my life! The planet needs more people like you! Thank you alot, because of you i understand so much!
how can I get the notes. It states unable to render code on Github
@@dhawalgore9338 try to download it and then open.
Awesome Summary this is one and only channel where I see the clear packet of necessary data outlines .. Hats off 😊
Bro learning fully this video enough for statistics in data science?? Plz give reply 🙂 for this question bro❤
Pretty much enough bro
Thanks from my bottom of the heart
Amazing video, thanks for cutting the video and making it more seamless. Your efforts are really appreciated. Thanks for this video.
In z-score section.. u can add right table + 0.5 and then subtract it from 1.. it gives the same thing
What a great resource!
very informative and helpful video for revising all concept and also for interview purpose thank you krish sir for making for us this like beautiful video
I am learning a lot from your Chanel. Thank you so much.
This is great Stats Playlist . Thanks for making this Krish
3:58:00 : if sample mean = 110, then Z = 3.65 & it's not in our calculated UB & LB range (-1.96 to +1.96) So, we reject the Null Hypothesis & it improves the Intelligence as Z > UB.
Thanks, one of the best course i've ever watch
Superb,Thank you Krish for amazing content.👏
Really sir ,very awesome concepts dilevary in simple manner
You are a great teacher!
Sir you are an absolute legend. I dont have an interveiw but this was useful to understand the concepts. Thank you!
Thanks Mr. Krish! Greetings from Chile!
Much needed one
If you could do the same for SQL and python would be very beneficial 😊
this is better than FSDA live classes.
thank you so much sir, you explained everything very well
Thank you so much Krish today I just have finished your lectures
Sir, I am going to do bstat in isi but I had not statistics at 10+ 2 level. Your video has helped me to start my journey with statistics by making every concept crystal like clear.
I appreciate you Sir, best Statistics course ever!
This is very nice thank you Mr Krish
Hello, for those who are looking for notes, you can go to one of his original live videos and from the description box you can go to the course dashboard and from there you can get the notes for each day separately in the resources section.
Thankyou so much
The dashboard has crashed though
@@anupamabalanmenon4771 not this one's description, the original live one's
Please share the link here.
HELLO Anshuman
I tried finding it but didn't got any
can you please help me with the link
Thanks!
1:24:00 = Why sample variance is divided by n-1?
ua-cam.com/video/vGsRwB3TsiE/v-deo.html
Summary : Researchers found that using the denominator (n-1) in sample variance/standard deviation calculations provides estimates closer to the population variance/standard deviation in various types of sample data distributions (positively/negatively skewed). This correction is also known as Bessel's correction or degrees of freedom.
thanks, one of the only people who explain all of this clearly haha!
sir with due respect you have made a major mistake in P-value and significance value Hypothesis conclusion 4:48:39 -
A p-value less than 0.05 (typically ≤ 0.05) is statistically significant. It indicates strong evidence against the null hypothesis, as there is less than a 5% probability the null is correct (and the results are random). Therefore, we reject the null hypothesis, and accept the alternative hypothesis.
You did the opposite.
Overall the course was good but minor mistakes here and there. Thank you
Though you corrected it later.. But best is when editing you could just add * and add comment on video. Because while studying and taking notes with lecture it becomes a very bad experience to go back and correct all the wrong. The thought process needs to changed fully to understand it again..
true
true, that was annoying
Bro learning fully this video enough for statistics in data science?? Plz give reply 🙂 for this question bro❤
True i was also confused and stuck for half an hour because he only contradict himself .
This helped me to clear my interview today. Thanks a ton for the statistics crash course.
What company and what role bro?
Yeah pls update will be useful for us
@@arunabhkumar6501 It's for the Data analyst role wayfair company, Question was how to handle the missing values & outliers in dimension and measure.
@@bapupatil9354hi did you have any prior experience, if not kindly share how did you apply for the job because I am fresher & trying to apply for job but not getting any calls as well as emails kindly support
@@Dineshhhh131 where are you applying
have you got the job
This is so needed 1.Consise 2.Detailed 3.No idiot is asking unnecessary questions
GREAT
You the man Krish!!
If p_value is less than 0.5 then we should reject the null hypothesis(Ho).
correct
On the measures of central tendency, I noticed, you kept referring to the median as the mode. Those are two different measures. Luckily you corrected it down the line. But overall, perfect summary of statistics.
Guru ji mst..... Teaching and depth
First of all thanks for producing such a useful and insightful video on Statistics.
Now my question is about exit poll results (almost all are failing). What I infer from that is:
1) Samples are biased- As they claim of random sampling but the samples are biased (gender biased, community biased, wrong answering biased).
2) Sample size- owing to humongous population of our country its quite impossible to collect even considerable sample data from all types of populations.
3) Biasing in result predictions- as can be seen all analysis of exit poll is agenda driven that is party specific. (Human bias)
4) Collection techniques- as technology progresses still these companies rely on old conventional way of inferring the results. Most of them rely on structured data or on survey reports but in todays Data driven world unstructured data can predict better results which all companies are lagging.
i have found this tutorial to be the best so far , A big thank you
Thank You for this video. I learnt a lot
2:08:07 , we can consider this approach , where the area between range 100 till 145 (right part) to be 50 % as the guassian distribution is generally a symmetric one.
and we also know , area between 85 to 115 is 68 % as it the between -1 SD to +1SD as per 68% 95% 99.7% rule , there fore area between 85 to 100 would 34 (68/2) due to its symmetry.
now on adding area between (85-100 and 100-115), we get 50+34 = 84 .
84% represents the region of people having IQ more than 85.
to find less than 85 we need to subract 100 from it, i.e. 100-84 = 16%
Thank you so much ..for uploading this video ..stats video ....😊
Thank you for the video. Just want to point out on P-value less than alpha , we reject the null hypothesis. 4:24:10 Kris fixed it later in the video.
exactly
you are right
fantastics course...Hats off sir
world's beast video ever on UA-cam about statistics
Thanks you sir! greetings from Chile!
thanks a lot sir for these efforts..truely a gem
Krish jaani tu cheeta hai.... Itne complex topics itni aasani se samjha diye.. Lots of love and respect from Pakistan
amazing lecture!
Nice one Mr Krish I love you
Hey Krish! Very much impressed with your videos. Currently i'm going through DS science course. and these videos are the life saving
for me.
Nice summary -well explained. !
Mam learning fully this video enough for statistics in data science??😊
@@itz_satya_3yes i believe
Thank you for making our life easier ....
God Bless you! 😀
02:24:31 In my case it will show empty list of outliers because 100, 110, and 115 are not extreme enough to be classified as outliers with a threshold of 3. So i change the threshold 3 to 2 this way it will work. My dataset is : dataset = [50, 52, 48, 47, 51, 49, 53, 45, 46, 54, 55, 50, 49, 52, 47, 48, 50, 51, 46, 53,100,110,115]
Watched the video in a day and a half, to revise stats that I had last studied in engineering days, helped me a lot cracking 3 quant interviews ending up in a 50% higher pay quant researcher role. Thanks brother. ❤
Can you please upload video on how to read data in excel using statistics like various percentiles? What insights do we get it ?
Thank you for the video.
Hi sir recently i ve seen this video !! So amazing and helpful for my data science carrer !!
Now i saw in other videos that in hypothesis testing ,
the type 1 error is false positive and type 2 error is false negative
Which book I can refer to learn all these statistics concepts you taught. It ll be great yo have some reference
Previously i had sawn a lot of vedios sir no vedio given me the proper understanding .Thanks a lot sir not only understandig,the vedio also given me with the examples of the interview questions
thanks thanks a lot sir🥰🥰🥰
Hi Krish.. could you please tell us something about chartered data scientist accreditation and how useful it is
Thank You Sir.
Hello Sir, in chi square census example ,why we need to choose the 0.05 in the chi square table ? actually it is two tail test right! so we need to see the value of 0.025 in the table with respect to df = 2 right ?
Yes u r right..Ho jata hai kabhi kabhi
Thank you sir from deep of our hearts
Your beautiful student
must video amazing video sir 😍😍😍
4:48:49 if p-values is less than alpha, then we need to reject the null hypothesis. This needs to be corrected in the video. Let me know if I missed something. It was a great explanation overall. Thanks
Later in the video, correction has been made. Thanks
Nice video 😊 and good explanation for each concept
thank you for the video. I have a subject named Introduction to Data Science using python and has statistic in my syllabus, as it was only introduction, i have completed the video till t - test and i have basic idea about covarience and correlation. My exam is tomorrow let's see if i pass.
Thank you so much for teaching
Exactly this, he got it very confused in the video. For other folks who are confused:
at 95% confidence, alpha = 0.05,
at 90% confidence, alpha = 0.1, this is within the confidence interval if alpha = 0.05
Looking at this, if alpha is 0.05, then a value > 0.05 (cuz 0.1 > 0.05) will fall within the confidence interval.
So, p>alpha implies that it lies within the confidence interval, so we accept the null hypotheses.
p REJECT THE NULL HYPOTHESIS
p > alpha (domain expert will tell this values) ---> ACCEPT THE NULL HYPOTHESIS
5hr:3min, Each tail will have 0.00889 and the middle region will have 0.9822, Z value gives the region below 2.307 which also includes area less than -2.307. So we need to subtract that tail value to get the middle region.
Bro learning fully this video enough for statistics in data science?? Plz give reply 🙂 for this question bro❤
I agree with with u I also think the same way bcoz while finding out the z score we r getting the area from +2.307 till -ve end not till -2.307, so i guess it should be each tail as 0.00889 which means P-value is 0.00889*2 and the middle region will be 1-(0.00889*2). But with all due respect, hats-off to u Krish sir for making stats soo simple to under. Because of u only we are able to spot even these minor things
Thank you, Krish!
Thank you very much sir
Hi Krish Sir,
Thanks for the amazing informative video. In type 1 and type 2 error while explaining confusion matrix TN should be type 2 error. Earlier you wrote correctly, but later you marked FN as type 2 error which is incorrect.
Ya, why no one pointed it out
I got confused. Chatgpt says FN is type-2 error
FP is type 1 and FN is type 2 error
Thanks a lot sir...
thank you so much sir 🙏 for this concept
dear sir,
with due respect i want to ask something
in case when we use trim mean and normal mean to find outlier and triimmean value is very close to normal mean ,it is said that outlier is not present in that particular table but when we check with quartile method there exist a outliers in the given table