Machine Learning Roadmap for Beginners ( If you have a Non-CSE background like me😉)

So, you are a non-CSE background student like me and trying hard to switch to machine learning/Data science field because you saw how world is moving towards AI. To be honest, there is nothing wrong with your thinking.

Back in my second year of university, I also realized that there were very few opportunities for the subject that I was studying, and I could not have the opportunities, flexibility, and, of course, salaries that I wanted in the field that I was studying. So I started to explore other fields that have a bright future and opportunities.

I decided to move to the data science field, finished a 5-month online course and now I doing my last-year thesis using my Machine Learning skills that I have learned so far.

Let's be honest, data science/ML is a combination of coding, statistics, math, communication, and understanding business problems. If you are preparing for a job right now, then you have to learn both soft and hard skills. But if you are a student and want to start learning ML right now, my suggestion is to focus on learning the hard skills (coding, math, doing projects, and business theory).

I will be writing this blog for students like me who are planning to learn ML in addition to their academic study and apply this skill in his field and after graduation, he can move to the DS/ML field completely.

So let's start!

Stage-1) Coding:
You need to start with a programming language. Python and R are the two most popular language in data science field.you can go with any of these.But my suggestion is to go with Python. Becouse Python is really easy to learn, and there are things that cannot be done with it: web development, app development, and data analysis. Python is like the One Man Army! I totally love Python.

• Topics
Variables, Numbers, and Strings
Lists, Dictionaries, Sets, and Tuples
o If condition, for loop
Functions, Lambda Functions
Modules (pip install)
o Read and write files
Exception handling
Classes, Objects

these basic topics are enough for beginees and move to next stage.

Stage-2) Data Analysis and Visualization:
Numpy, Pandas, Data Visualization Libraries (Matplotlib and Seaborn)

Stage-3) Math and Statistics For Machine Learning:

Topics to Learn
▪ Basics: Descriptive vs inferential statistics, continuous vs discrete data,
nominal vs ordinal data
▪ Linear Algebra: Vectors, Metrices, Eigenvalues and Eigenvectors
▪ Calculus: Basics of integral and differential calculus
▪ Basic plots: Histograms, pie charts, bar charts, scatter plots, etc.
▪ Measures of central tendency: mean, median, mode
▪ Measures of dispersion: variance, standard deviation
▪ Probability basics
▪ Distributions: Normal distribution
▪ Correlation and covariance
▪ Central limit theorem
▪ Hypothesis testing: p value, confidence interval, type 1 vs type 2 error, Z test

Stage-4) Exploratory Data Analysis (EDA):
Time for projects! use the skills that you have learned so far and do some Data Analysis projects. EDA is extremely important in machine learning, as it is necessary for data preprocessing part. For datasets, go to Kaggle.

stage-5) Machine Learning:
• Machine Learning: Preprocessing
Handling NA values, outlier treatment, and data normalization
One hot encoding, label encoding
Feature engineering
Train test split
Cross-validation
• Machine Learning: Model Building
Types of ML: Supervised, Unsupervised
Supervised: Regression vs Classification
Linear models
▪ Linear regression, logistic regression
▪ Gradient descent
Nonlinear models (tree-based models)
▪ Decision tree
▪ Random forest
▪ XGBoost
o Model evaluation
▪ Regression: Mean Squared Error, Mean Absolute Error, MAPE
▪ Classification: Accuracy, Precision-Recall, F1 Score, ROC Curve,
Confusion matrix
o Hyperparameter Tuning: GridSearchCV, RandomSearchCV
Unsupervised: K means Hierarchical clustering, Dimensionality reduction
(PCA)
Do at least 2 end-t0o-end Machine learning projects and deployment.

For delopyment you can use Stremelit and for advance you must learn Python web development frameworks- Flust,FastAPI,Django.I suggest FastAPI.

That's it! Now you have enough skill to start doing projects and learn further on your own.Some of you may ask, "Shemanto, you didn't write about t SQL, MLOp and other advanced Python concepts.". You are right, I didn't write then on purpose.

You see, ML skill are specific knowledge.You can not learn it just by watching tutorial.You have to spend less time in consuming information and more time
o Digesting
o Implementing
o Sharing

if you keep going like this, you can find out other stuffs on your own.I just shared 5 stages that are easy to start, and I think that is enough for a bigess to tell everyone, "I have machine learning skills.".

Keep Learning!