Harbola DataScience

Answer, Read and Post Coding related Questions!

What are the ‘Important Machine Learning Topics’ to Start Building Amazing Projects

If you are reading this post there are high chances of you knowing how vast is Machine Learning and achieve expertise in it required years of hands on experience on real world projects. Learning all the theoritical concepts at once can a huge blunder as there are tons of topics and skills required in Machine Learning such as

Maths - statistics, Linear Algebra, Probability, and Calculus
algorithms (Supervised Learning)-      
    Linear Regression
    Logistic Regression
    Decision Trees
    K-Nearest Neighbors
    Naive Bayes
    Support Vector Machines
    Ensemble Learning Techniques
Semi-Supervised Learning
Unsupervised Learning
Reinforcement Learning
Neural Network or Artificial Neural Network (ANN)

Each of them is a huge field in itself. And interesting part is that you need not to be master at every field. Having a little knowledge about everything and knowing everything about few concepts is everything you need to know. Following Pareto Principle (80-20 rule) works best in areas where there is huge syllabus.

Machine learning is making computer system smart enough using lot of data so to take bussiness decision, insights, product development, prediction, recommender system applying algorithm and calculations.

What to know for Machine Learning project

Machine learning is a very interesting sub-field of datascience in which we play around with data using some tools to clean data and apply algorithms to predict future/unknown outcomes. Python and R are some of the most popular programming languages used for ML.

To build a project around machine learning these important topics are enough,

pandas for data manupulation,
numpy for numerical calculations,
matplotlib to plot graphs,
sklearn to apply prebuilt algorithms,
concept of train-test splitting,
error analysis (accuracy)

that’s it doing these few topics from basic will take you 1-2 week to build your first ml project. There are lot of dataset present on which you can fit you model (algorithm) and predict the values. and check the accuracy subtracting the pridicted value from actual values.

Application –

google colab notebook IDE

Simply using some libraries and packages you can easily bring your dataset into workspace

data cleaning

using pre-build methods of pandas and numpy you can easily do most (99.9%) of the task without writing manual code from scratch.

pandas profiling – build reports

you can build data report in which you will get most of the insight about the data at one place in a html webpage. with a single run.

using seaborn library to plot graph

this not as that good graph but ya its easy to build graph using python packages.

Most Important part

spliting data into two parts

we need 2 set of data. one to train our model and another to test on. By using text_size attribute we can specify the ratio of test and train set.

model creation – random forest algorithm

sklearn library has made it so easy to apply algo on our dataset. By simple .fit() command we can train our model. The algorithm is pre-written. At beginner level you don’t need to go that deep.

You can see that using random forest algorithm we are getting 93% accuracy which is pretty high. And that’s it.

By .predict() method you can predict the future outcome of upcoming values.

Similarly using these few concepts and libraries you can easily build your model and start your machine learning journey. And leave a mail @ chitranshuharbola@gmail.com if you are interested in all of these things. We are doing many interesting things and you can be part of it 🙂

Hope I would have solve your question of "what are important machine learning topics required for projects" And please post your opinion on this topics on comments down below and some other doubts and question related to datascience or machine learning. :)

Chitranshu Harbola

Self taught programmer, Web Developer and an aspiring Machine learning engineer cum Data Science student

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to top