Author: Tom
-
Types of Statistics Descriptive, Predictive & Prescriptive
Descriptive, predictive, and prescriptive are applicable to both statistics and machine learning. Descriptive problem: predictive problem: Prescriptive problem:
-
Statistics vs Machine learning
Statistics Machine learning Problem > Data > Solution Data > Solution w/o problem Question is cheap, Data is expensive Data is cheap, Question is expensive Youre paid for collecting the data Youre paid for asking the right question
-
Standard Normal Distribution
Every distribution can be standardized Assuming μ as the mean and σ2 as the variance, the standardization is a process of transforming this variable to 0 and 1 respectively. Standardization: ~(μ,σ2) -> ~(0,1) When we standardize a normal distribution, the result is a standard normal distribution. If we shift the mean by μ and standard…
-
Normal Distribution
The normal distribution is the most common type of distribution. It is also referred to statistically as Gaussian Distribution. Many people call it the bell curve because of its shape. it is symmetrical and its mean = median = mode. It has no skew as its perfectly centered around its mean. Keeping standard deviation constant…
-
Distribution
In statistics, whenever we use the term distribution, we usually mean a probability distribution. A distribution is a function that shows the possible values for a variable and how often they occur. The distribution of an event consists not only of the input values that can be observed but is made up of all the…
-
Measures of Relations Between Variables: Correlation
Linear Correlation Coefficient Correlation adjusts covariance, so that the relationship between the two variables becomes easy and intuitive to interpret. Note: -1 <= correlation coefficient <= 1 Correlation coefficient of 1 implies, the entire variability of one variable is explained by the other. they are moving in same direction. Correlation coefficient of 0 implies, the…
-
Measures of Relations Between Variables – Covariance
Covariance If two variables are correlated (eg: house size and price), the main statistic used to measure this correlation between two variables is called covariance. Covariance gives a sense of direction in which the two variables are moving: one problem with covariance is the values can be of completely different scales for each dataset (eg:…
-
Measures of Variability – Standard Deviation & Coefficient of variation
Standard Deviation In most analysis we perform, the standard deviation will be much more meaningful than variance. Standard deviation is the most common measure of variability for a single dataset. Standard deviation is the square root of variance. for both population and sample. Coefficient of variation (cv) Comparing Standard deviation between two datasets is meaningless.…
-
Measures of Variability – Variance
Variance Variance measures the dispersion of the set of data points around their mean. Example: Population: 1, 2, 3, 4, 5 The sample variance is higher than the population variance for the same data in the above example because: by calculating the sample variance of the population, we are telling that dataset 1,2,3,4&5 are sample…
-
Measures of Asymmetry
Skewness The most commonly used tool to measure asymmetry is Skewness Skewness indicates whether the observations in a dataset are concentrated on one side. There are three types of skewness: positive/right skew, zero skew & negative/left skew. The formula for skewness is not important as it will be calculated by the program itself. The formula…