prerequisites and why Math for machine learning

Presented as a talk at CHAI(coffee with AI) Hyderabad #1 hosted by learn.machinelearning Instagram page community.

I talked on what are the prerequisites and why mathematics is important for machine learning and how to get started.

Transcript

1. Welcome CHAI Hyderabad #1
2. Topics • What are the prerequisites? • Do I need to learn MATH? • What approach to follow?
3. prerequisites Programming skills Data analysis Probability & statistics Linear algebra Calculus
4. Why Python??? • Simple and consistent • Easy to learn • Easily readable syntax • Fewer lines of code • Understandable by humans • Helps you to focus on problem solving • Write reliable code
5. Why Python??? • A great library ecosystem • Hundreds of libraries • Save time • well-structured and well-tested • Examples • Scikit learn • Pandas • Keras • TensorFlow • Etc...
6. Why Python??? • Flexibility • OOP or scripting • No need to recompile the source code • Can combine Python and other languages • Different programming styles • Flexibility factor decreases the possibility of errors
7. Why Python??? • Community • Platform independence • Good visualization options • Many more
8. Why Data analysis??? (You can answer unknown questions) 80% of the work will be with the data Getting data Exploring & Visualizing data Cleaning data
9. Why Data analysis??? • Exploring Data • Understand the data first and know your goal • It gives hints for Data Cleaning • It give ideas for feature engineering • Data size? No of features? Type of a feature? Target variable? • Data distribution plots • Correlation plots
10. Why Data analysis??? • Data cleaning • One of the typical challenge is feeding the right data • Better data beats fancier algorithms • garbage in gets you garbage out. • Goal of data cleaning is to identify and remove errors like • Missing data • Outliers • Bad data and duplicates • Irrelevant features • Standardization
11. Why Statistics and Probability??? Probabilitydeals with predicting the likelihood of future events, while statistics involves the analysis of the frequency of past events. Collection of tools used to answer important questions about data Inferential statistics and descriptive statistical methods Descriptive statistics will help to gain knowledge from the raw data Inferential statistics help to draw conclusions from the data Difference between statistical models and machine learning
12. Why Linear algebra??? • Helps you to understand the notations and formulas • Paper to code • Matrix factorization • Linear algebra = linear equations = matrices + vectors
13. How to get started? • Adjust Mindset • Beginner • Experienced • Academic • Industry