How to start a Machine-Learning journey
Amazon Alexa, Microsoft Cortana, and Google Home really amaze a lot of people. Their curiosity really lead some of them to know more about how are they built?.In googling for more information about these products led them to machine learning(ML) tutorials, data science(DS) tutorials etc. Now they want to start it but don’t know which concept to start first? how to start? etc.Proper path to learning is the key to learn fast and learn good.
In this article, We will talk about how to approach the correct learning path of ML and DS and which concepts you should refine before going on the boat of ML.Before starting the journey ask this question?
Why should I study ML? and what benefit will I give to society after learning this?
This question is very important because you need a motivating force to keep you on track in this hard voyage.You should make small goals ultimately taking you to a big goal. Procedure of learning is as follows:
Python: This is a programming language that you need to learn first.For ML there is no specific language but most of the work in ML is going on in python language and it is easy to learn. In python focus more on data structures, libraries like numpy, matplotlib ,pandas etc.
Statistics: Learn the basics of stats and probability like conditional probability,probability density function, cumulative distribution function, mean, median, mode, standard deviation, variance, Normalization, Standardization, Correlation, Co-variance and different type of distributions etc.
Linear Algebra: Learn the basics of Linear Algebra more about vectors, Dot product, angle between two vectors, projection, Unit vector, Equation of line, Distance of point, Hyper cube, Hyper Cuboid etc.
Data Science Basics: Learn the basics of data science more about data acquisition, Exploratory Data Analysis, Data wrangling, Data manipulation, Accessing important feature, Data Plotting(pair plots, box plots, violin plots)etc and then data visualization.
Dimensionality Reduction Techniques: Learn techniques like Principal component analysis and T-SNE (t-distributed stochastic neighbor embedding).As data in ML may have thousands of dimensions so in order to visualize it, dimension reduction is necessary.Apply these techniques on MNIST dataset to properly grasp the knowledge.
Machine Learning Algorithms: Start with supervised learning and learn ML algos like KNN, Naive Bayes, Linear Regression, Logistic Regression, Support vector machines, Random Forrest, Ensemble Models.These algorithm are applied in python using Scikit Learn library.Go to the website and read all documentation related to classification, Regression and clustering(unsupervised learning).
Machine Learning is a hard journey with a lot burn outs but dedication and schedule planning can keep you on track. Learn it with focus and practice it daily.
