STEP 1
Curriculum: this is the first month of our Machine Intelligence course, let’s learn together and grow together
I am dividing the subject in small bits, so that it is easier to conquer for amateurs, non stats/ maths background people to cover.
DAY 1 : Basic Statistics
Variables
Sample, population,
Deterministic and stochastic variables,
Descriptive statistics (moments about point, mean, median, mode, percentiles, Inter Quartile Range
Independence, covariance, Correlation, cause and effect
Probability basics(addition, multiplication, union, intersection, conditional probability, Bayes theorem, )
PDFs, CDFs, Discrete and continuous variables Distributions(Discrete Uniform, Bernoulli, Binomial, Poisson, Continuous Uniform, Gamma, exponential, Chi square, Gaussian, Standard Normal, Lognormal, Multi Gaussian), Joint pdfs
Parameters, Estimation of parameters, Techniques of estimating parameters (Method of Moments, Method of Maximum Likelihood Estimate)
Hypothesis Testing, p- value and significance levels,
Different tests: t-test, ANOVA, chi square test, f test for variance
Bayesian vs frequentist, Generative vs Discriminative
DAY 2: Basic Mathematics
Sets, Relations, functions, image of function
Calculus: Differentiation, Partial Derivatives, Jacobian Matrix, Hessian Matrix, beta and gamma functions, Integration, Leibnitz integral rule, Partial Differential Equations
Limit, continuity, Differentiability, Mean Value Theorem, Indeterminate Form, L’ Hopital’s Rule
Product and Chain Rule of infinite series,
Linear Algebra: Matrix Addition, Matrix multiplication, scalar multiplication, linear transformation, transpose, conjugate, rank, determinate, inverse, Pseudo inverse, inner and outer product, matrix factorization concepts, Eigen values, trace of matrix, Eigen Vectors, diagonalization, singular value decomposition, positive definite matrix
Matrix Algebra for Engineers: Jeffrey Chasnov
Convex problems, Convex Optimization, point of Optimization, Optimization Techniques, ( Methods of finding roots), minimization, maximization, inflexion points, , Linear programming problems, converting duals, loss function, Lagrange multipliers
Gradient Descent, Stochastic Gradient descent, Batch Gradient Descent
Playlists of Maths/stats : Khan Academy
DAY 3: Basic Python
Getting Started : About python, python syntax, variables
Operators : Arithmetic, Bitwise, Assignment, Comparison, Logical, Identity, Membership
Data Types : Numbers(including Booleans), Strings, Lists, Tuples, Sets, Dictionaries, Type Casting
Control Flow : if, for, while, try, break, continue
Functions : Function definition(def and return), Scopes of variables inside functions(local and global), Default argument values, Keyword arguments, Arbitrary Argument lists, Unpacking Argument Lists(* and **), Lambda Functions, Documentation strings, Function Annotation, Function decorators
Classes : Class definition, Class object, Instance Object, Method Objects, Class and instance variables, Inheritance, Multiple Inheritance
Modules : Modules, Packages
Errors and Exceptions : Handling exceptions, Raising exceptions
Files : Reading and writing (with statement), Methods of File objects
Additions : pass statement, Generators
Brief Tour of the standard Libraries : Serialization(json library), File Wildcards, String Pattern Matching, Mathematics(math, random, statistics), collections, Dates and Times, Data Compression
DAY 4: Basics of Machine Learning
Supervised(regression, classification), unsupervised(clustering, association), ai, deep learning,
Fitting data, quantitative models(explanatory vs time series- (AKnotes)), Underfitting, Overfitting,
Bias and Variance
Types of data, Getting data in proper shape and size, training and test, feature Transformation, missing data, EDA
Multicollinearity, Cross Validation, Curse of Dimensionality, Hyperparameter, Gradient Descent, Stochastic Gradient Descent,
DAY 5: Python Implementation of EDA
Two sklearn datasets manipulation and massaging,
getting into right format
sklearn, Seaborn, matplotlib
DAY 6: First algorithm of Regression : Linear Regression
the math, Ordinary Least Square calculation with two variables, OLS is BLUE(linear, Unbiased, Least Variance Estimate), extending to multivariate, measuring the accuracy like R sq,, p-value etc., Back Elimination(https://www.javatpoint.com/backward-elimination-in-machine-learning) , Polynomial Regression
Regularization, Extension into Ridge regression, lasso regression
DAY 7: Python code of linear regression(code sharing for from scratch)
sklearn implementation, cross validation and linear regression
reading results, visualisation
DAY 8: Logistic regression math, calculation with two variables, loss function, extending to multivariate, softmax regression, math, multiclass classification,
DAY 9: Implementation of Logistic regression( share code from scratch)
Extended Materials:
StatQuest , Sebastian Raschka , Great Learning Statistics , Bayesian vs frequentist, Generative vs Discriminative , Cassie Kozyrkov , Statquest Machine learning Playlist
