Master statistics & machine learning: intuition, math, code

Master statistics & machine learning: intuition, math, code

English | MP4 | AVC 1280×720 | AAC 44KHz 2ch | 224 lectures (38h 13m) | 12.8 GB

A rigorous and engaging deep-dive into statistics and machine-learning, with hands-on applications in Python and MATLAB.

Statistics and probability control your life. I don’t just mean What YouTube’s algorithm recommends you to watch next, and I don’t just mean the chance of meeting your future significant other in class or at a bar. Human behavior, single-cell organisms, Earthquakes, the stock market, whether it will snow in the first week of December, and countless other phenomena are probabilistic and statistical. Even the very nature of the most fundamental deep structure of the universe is governed by probability and statistics.

You need to understand statistics.

Nearly all areas of human civilization are incorporating code and numerical computations. This means that many jobs and areas of study are based on applications of statistical and machine-learning techniques in programming languages like Python and MATLAB. This is often called ‘data science’ and is an increasingly important topic. Statistics and machine learning are also fundamental to artificial intelligence (AI) and business intelligence.

If you want to make yourself a future-proof employee, employer, data scientist, or researcher in any technical field — ranging from data scientist to engineering to research scientist to deep learning modeler — you’ll need to know statistics and machine-learning. And you’ll need to know how to implement concepts like probability theory and confidence intervals, k-means clustering and PCA, Spearman correlation and logistic regression, in computer languages like Python or MATLAB.

There are six reasons why you should take this course:

This course covers everything you need to understand the fundamentals of statistics, machine learning, and data science, from bar plots to ANOVAs, regression to k-means, t-test to non-parametric permutation testing.

After completing this course, you will be able to understand a wide range of statistical and machine-learning analyses, even specific advanced methods that aren’t taught here. That’s because you will learn the foundations upon which advanced methods are build.

This course balances mathematical rigor with intuitive explanations, and hands-on explorations in code.

Enrolling in the course gives you access to the Q&A, in which I actively participate every day.

I’ve been studying, developing, and teaching statistics for 20 years, and I’m, like, really great at math.

What you’ll learn

  • Descriptive statistics (mean, variance, etc)
  • Inferential statistics
  • T-tests, correlation, ANOVA, regression, clustering
  • The math behind the “black box” statistical methods
  • How to implement statistical methods in code
  • How to interpret statistics correctly and avoid common misunderstandings
  • Coding techniques in Python and MATLAB/Octave
  • Machine learning methods like clustering, predictive analysis, classification, and data cleaning
Table of Contents

1 [Important] Getting the most out of this course
2 About using MATLAB or Python
3 Statistics guessing game
4 Using the Q&A forum
5 (optional) Entering time-stamped notes in the Udemy video player

Math prerequisites
6 Should you memorize statistical formulas
7 Arithmetic and exponents
8 Scientific notation
9 Summation notation
10 Absolute value
11 Natural exponent and logarithm
12 The logistic function
13 Rank and tied-rank

IMPORTANT Download course materials
14 Download materials for the entire course

What are (is ) data
15 Is data singular or plural
16 Where do data come from and what do they mean
17 Types of data categorical, numerical, etc
18 Code representing types of data on computers
19 Sample vs. population data
20 Samples, case reports, and anecdotes
21 The ethics of making up data

Visualizing data
22 Bar plots
23 Code bar plots
24 Box-and-whisker plots
25 Code box plots
26 Unsupervised learning Boxplots of normal and uniform noise
27 Histograms
28 Code histograms
29 Unsupervised learning Histogram proportion
30 Pie charts
31 Code pie charts
32 When to use lines instead of bars
33 Linear vs. logarithmic axis scaling
34 Code line plots
35 Unsupervised learning log-scaled plots

Descriptive statistics
36 Descriptive vs. inferential statistics
37 Accuracy, precision, resolution
38 Data distributions
39 Code data from different distributions
40 Unsupervised learning histograms of distributions
41 The beauty and simplicity of Normal
42 Measures of central tendency (mean)
43 Measures of central tendency (median, mode)
44 Code computing central tendency
45 Unsupervised learning central tendencies with outliers
46 Measures of dispersion (variance, standard deviation)
47 Code Computing dispersion
48 Interquartile range (IQR)
49 Code IQR
50 QQ plots
51 Code QQ plots
52 Statistical moments
53 Histograms part 2 Number of bins
54 Code Histogram bins
55 Violin plots
56 Code violin plots
57 Unsupervised learning asymmetric violin plots
58 Shannon entropy
59 Code entropy
60 Unsupervised learning entropy and number of bins

Data normalizations and outliers
61 Garbage in, garbage out (GIGO)
62 Z-score standardization
63 Code z-score
64 Min-max scaling
65 Code min-max scaling
66 Unsupervised learning Invert the min-max scaling
67 What are outliers and why are they dangerous
68 Removing outliers z-score method
69 The modified z-score method
70 Code z-score for outlier removal
71 Unsupervised learning z vs. modified-z
72 Multivariate outlier detection
73 Code Euclidean distance for outlier removal
74 Removing outliers by data trimming
75 Code Data trimming to remove outliers
76 Non-parametric solutions to outliers
77 Nonlinear data transformations
78 An outlier lecture on personal accountability

Probability theory
79 What is probability
80 Probability vs. proportion
81 Computing probabilities
82 Code compute probabilities
83 Probability and odds
84 Unsupervised learning probabilities of odds-space
85 Probability mass vs. density
86 Code compute probability mass functions
87 Cumulative distribution functions
88 Code cdfs and pdfs
89 Unsupervised learning cdf’s for various distributions
90 Creating sample estimate distributions
91 Monte Carlo sampling
92 Sampling variability, noise, and other annoyances
93 Code sampling variability
94 Expected value
95 Conditional probability
96 Code conditional probabilities
97 Tree diagrams for conditional probabilities
98 The Law of Large Numbers
99 Code Law of Large Numbers in action
100 The Central Limit Theorem
101 Code the CLT in action
102 Unsupervised learning Averaging pairs of numbers

Hypothesis testing
103 IVs, DVs, models, and other stats lingo
104 What is an hypothesis and how do you specify one
105 Sample distributions under null and alternative hypotheses
106 P-values definition, tails, and misinterpretations
107 P-z combinations that you should memorize
108 Degrees of freedom
109 Type 1 and Type 2 errors
110 Parametric vs. non-parametric tests
111 Multiple comparisons and Bonferroni correction
112 Statistical vs. theoretical vs. clinical significance
113 Cross-validation
114 Statistical significance vs. classification accuracy

The t-test family
115 Purpose and interpretation of the t-test
116 One-sample t-test
117 Code One-sample t-test
118 Unsupervised learning The role of variance
119 Two-samples t-test
120 Code Two-samples t-test
121 Unsupervised learning Importance of N for t-test
122 Wilcoxon signed-rank (nonparametric t-test)
123 Code Signed-rank test
124 Mann-Whitney U test (nonparametric t-test)
125 Code Mann-Whitney U test
126 Permutation testing for t-test significance
127 Code permutation testing
128 Unsupervised learning How many permutations

Confidence intervals on parameters
129 What are confidence intervals and why do we need them
130 Computing confidence intervals via formula
131 Code compute confidence intervals by formula
132 Confidence intervals via bootstrapping (resampling)
133 Code bootstrapping confidence intervals
134 Unsupervised learning Confidence intervals for variance
135 Misconceptions about confidence intervals

136 Motivation and description of correlation
137 Covariance and correlation formulas
138 Code correlation coefficient
139 Code Simulate data with specified correlation
140 Correlation matrix
141 Code correlation matrix
142 Unsupervised learning average correlation matrices
143 Unsupervised learning correlation to covariance matrix
144 Partial correlation
145 Code partial correlation
146 The problem with Pearson
147 Nonparametric correlation Spearman rank
148 Fisher-Z transformation for correlations
149 Code Spearman correlation and Fisher-Z
150 Unsupervised learning Spearman correlation
151 Unsupervised learning confidence interval on correlation
152 Kendall’s correlation for ordinal data
153 Code Kendall correlation
154 Unsupervised learning Does Kendall vs. Pearson matter
155 The subgroups correlation paradox
156 Cosine similarity
157 Code Cosine similarity vs. Pearson correlation

Analysis of Variance (ANOVA)
158 ANOVA intro, part1
159 ANOVA intro, part 2
160 Sum of squares
161 The F-test and the ANOVA table
162 The omnibus F-test and post-hoc comparisons
163 The two-way ANOVA
164 One-way ANOVA example
165 Code One-way ANOVA (independent samples)
166 Code One-way repeated-measures ANOVA
167 Two-way ANOVA example
168 Code Two-way mixed ANOVA

169 Introduction to GLM regression
170 Least-squares solution to the GLM
171 Evaluating regression models R2 and F
172 Simple regression
173 Code simple regression
174 Unsupervised learning Compute R2 and F
175 Multiple regression
176 Standardizing regression coefficients
177 Code Multiple regression
178 Polynomial regression models
179 Code polynomial modeling
180 Unsupervised learning Polynomial design matrix
181 Logistic regression
182 Code Logistic regression
183 Under- and over-fitting
184 Unsupervised learning Overfit data
185 Comparing nested models
186 What to do about missing data

Statistical power and sample sizes
187 What is statistical power and why is it important
188 Estimating statistical power and sample size
189 Compute power and sample size using G Power

Clustering and dimension-reduction
190 K-means clustering
191 Code k-means clustering
192 Unsupervised learning K-means and normalization
193 Unsupervised learning K-means on a Gauss blur
194 Clustering via dbscan
195 Code dbscan
196 Unsupervised learning dbscan vs. k-means
197 K-nearest neighbor classification
198 Code KNN
199 Principal components analysis (PCA)
200 Code PCA
201 Unsupervised learning K-means on PC data
202 Independent components analysis (ICA)
203 Code ICA

Signal detection theory
204 The two perspectives of the world
205 d-prime
206 Code d-prime
207 Response bias
208 Code Response bias
209 F-score
210 Receiver operating characteristics (ROC)
211 Code ROC curves
212 Unsupervised learning Make this plot look nicer

A real-world data journey
213 Note about the code for this section
214 Introduction
215 MATLAB Import and clean the marriage data
216 MATLAB Import the divorce data
217 MATLAB More data visualizations
218 MATLAB Inferential statistics
219 Python Import and clean the marriage data
220 Python Import the divorce data
221 Python Inferential statistics
222 Take-home messages

Bonus section
223 About deep learning
224 Bonus content