Machine Learning with Python: Basics to Advanced Analytics

Explore data-driven insights with Python across beginner to advanced machine learning concepts, from data collection and statistics to real-life applications like sentiment analysis and recommendation engines.

Explore analytics in the big data ecosystem, tracing the evolution from batch reporting to predictive analysis and content analysis, and learn how data-driven insights inform actionable decisions in machine learning.

Explore how big data shapes machine learning by examining volume, velocity, variety, and data sources from ERP, web, social, GPS, video, and platforms like Hadoop.

Explore how globalization and digital technology push organizations up the sophistication curve toward predictive analytics and cross-functional data insights.

Explore descriptive, predictive, and prescriptive analytics and data mining—from standard reports to optimization—and recognize that data is gold, converting to knowledge.

Learn how machine learning enables computers to learn from data, enabling real-world applications like recommendations, fraud detection, and pattern recognition as data and compute expand.

Compare supervised and unsupervised learning, covering tasks like classification, prediction, and clustering with techniques such as k-means, k-medoids, hierarchical clustering, and DBSCAN; distinguish structured and unstructured data.

Discover the fundamentals of sampling in machine learning, explaining why we sample, how to randomly select representative samples, and how they reveal properties of a larger population.

Define the target population and elements, use sampling units and a sampling frame to study resident males in India, and understand parameter, statistic, sampling error, and non-response bias.

Explore errors in observed data, including interviewer bias, respondent bias, and instrument bias with zero errors, and compare probability sampling (random, systematic, stratified) to non-probability sampling (convenience, judgment, quota, snowball).

Explore probabilistic sampling techniques, including random, systematic, and stratified sampling, and see how they reduce bias while selecting representative samples from a population.

Explore cluster sampling as a probabilistic technique used when a sampling frame is costly or unavailable, with examples, and compare it to non-probability methods like convenience, judgment, and quota sampling.

Identify the nature of data as qualitative or quantitative and explore their levels of measurement. Learn why selecting the right data type is essential for valid research and data science.

Explore qualitative data with nominal and ordinal levels. Differentiate quantitative data as continuous or discrete, and visualize with pie charts, bar charts, Pareto, and histograms for exploratory data analysis.

Master probability concepts for machine learning, including discrete and continuous random variables, distributions, events, conditional probability, Bayes theorem, and sample space for experiments.

Learn the relative frequency interpretation of probability through sample space, events, simple and compound events. Apply calculations from relative frequencies using real-world examples like service requests and telecom surveys.

Explore how probability models solve real-life problems by using joint probability of events. Learn how A and B intersect to show when both occur, with practical retail and travel examples.

Explore conditional probability and the formula p(a|b) = p(a ∩ b)/p(b), then apply it to ticket-resolution data and discuss lift in business analytics.

Explore conditional probability and define independence, showing when p(A|B) equals p(A) and when A and B multiply to p(A) p(B); illustrate with service tickets and bread and butter examples.

Apply the law of total probability for mutually exclusive and collectively exhaustive events to compute P(A) from P(A|B_j)P(B_j), then introduce Bayes theorem for inverse probability with practical examples.

Explore random variables as outcomes of a random process, linking sample space to numerical values. Distinguish discrete versus continuous variables with examples like coin toss, die rolls, and failure times.

Explore probability distributions as the complete description of a random variable, and contrast discrete and continuous cases using PMF and PDF that sum or integrate to one.

Explore cumulative probability distribution and the cumulative distribution function, with discrete, continuous, and mixed distributions, using a die example to illustrate probabilities, expectation, and variance.

Explore the Bernoulli distribution as a binary outcome with x=1 probability p and x=0 probability 1-p, revealing its mean p and variance p(1-p). Extend to the binomial distribution as the sum of independent Bernoulli trials.

Examine the binomial distribution using n and p, and compute the mean np and variance np(1-p). See how Gaussian (normal) distribution approximates binomial for large n and introduce geometric distribution.

Explore the geometric distribution's probability function and the Poisson distribution's arrival rate lambda for counts. Learn about mean and variance, and apply Chebyshev's inequality to bound probabilities in data.

Explore Chebyshev's inequality and its bounds on deviations from the mean, and contrast them with the normal (Gaussian) distribution, its probability density function, and mu and sigma.

Learn matrix algebra foundations, including scalars, vectors, and matrices, and master dimensions and the transpose operation. Explore matrix equality and square matrices as building blocks for computation.

Explore matrix operations by reviewing addition and subtraction of equal-order matrices, scalar multiplication, and the rule of conformability for matrix multiplication, with practical examples and intuition.

Explore real-life matrix multiplication across varying dimensions, observe how reversing order changes results, study diagonal, identity, and null matrices, transposes, determinants, and even principal component analysis basics.

Explore determinant properties, including det A equals det A transpose, sign changes from row and column swaps, and the role of rank and inverse in matrices.

Explore type I and type II errors, null and alternative hypotheses, and how alpha, beta, and power guide hypothesis testing decisions, including p-value and critical value approaches.

Learn the critical value approach to hypothesis testing, set null and alternative hypotheses, compute the t statistic for the population mean, and decide by alpha against the critical value.

Examine right, left, and two-tailed t tests with a 14 df sample, alpha 0.05, and mu=3. Determine rejection of the null via critical t values and the observed t statistic.

Master the p-value approach to hypothesis testing, comparing p to alpha, with four steps and null/alternative hypotheses and area under the curve.

Learn how p-values assess evidence against the null hypothesis, reject when p is less than alpha, and apply left-tail, right-tail, and two-tailed tests using both critical value and p-value approaches.

Conduct a one-sample right-tail hypothesis test on 25 Bernal hardness measurements to assess if the mean exceeds 170, and interpret the p-value, t-statistic, and 0.05 significance level.

Assess whether the vinca minor root extract lowers sunflower seeding height using a left-tailed t-test on 33 seeds, mean 13.66, p < 0.01, rejecting mu = 15.7.

Explore two-tailed hypothesis testing with null and alternate hypotheses, significance levels, p-values, and t-values, including rules to reject or fail to reject the null.

Learn to estimate population means and proportions using confidence intervals with a margin of error. Explore real examples, such as poll results and their bounds on public opinion.

Compute a t interval for the population mean using alpha, the t multiplier, and the standard error, then interpret the resulting confidence interval with an fvc example.

Review the normal distribution, contrast parametric and non-parametric data, explain standardization to a standard normal distribution, and apply the empirical rule to assess normality.

Assess how closely sample data follows a normal distribution with qualitative and quantitative methods, using q-q plots, Kolmogorov-Smirnov, and Shapiro-Wilk, and interpret normal probability plots within a 95% confidence band.

the lecture uses the Anderson-Darling test on ten weeks of expenditures, interprets a p-value of 0.742 as normality, and discusses outliers, transformations, and parametric versus non-parametric data.

Apply Box-Cox power transformation to find an optimal lambda that normalizes data, often via log transformation, and evaluate normality with tests like Anderson-Darling or Shapiro-Wilk.

Explore t tests for small samples, including one-sample and two-sample t tests, and compare the student's t distribution with the standard normal distribution to interpret results.

Explore one-sample and two-sample t tests using real-life data, formulate null and alternative hypotheses, compute t statistics and p-values, and interpret confidence intervals with practical Python and Minitab demos.

Compare two group means with a two-sample t test, interpret the t distribution, p values, and confidence intervals, and visualize differences using box plots and dot plots in hypothesis testing.

Learn chi square test of independence for two categorical variables, using contingency tables to compare observed and expected frequencies and interpret results with degrees of freedom and alpha.

Assess independence between gender and education level using chi-square tests at a 0.05 level of significance. Interpret contingency table results, p-values, and confidence levels to draw conclusions.

Explore the chi-square goodness-of-fit test, interpreting critical values and p-values to assess data fit. Use a 20%/80% resident vs non-resident example to apply a null hypothesis.

Compute and interpret a chi-squared goodness-of-fit test, compare observed vs expected frequencies, and assess independence in a contingency table using p-values and critical values.