Machine Learning is Fun!
A Simplest introduction to Beginners
what is Machine Learning?
What are the Libraries used in ML?
Where should we start?
How is deep learning related to ML?
.
.
And many more…
If you have the similar questions and searching for perfect answers then this article is for you. This guide is for anyone who is curious about machine learning but has no idea where to start.
Everything in my articles(Including this) will be in simple English and easier to understand.
Without discussing more lets get into the topic…:)
Coding is the only way to empower the unconscious mind. From starting ML from scratch to mastering ML, one should know how to code in python.
Also i suggest to learn R language, R is mainly used when the data analysis tasks require standalone computing or analysis on individual servers.
What is Machine Learning?
ML is just math. If you are really interested in math and you know how to code in python then this is your cup of tea.
Machine learning is an application of Artificial Intelligence(AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.
In simpler words, Machine Learning gives the ability to learn themselves using computers programs that can access the data.
You might have heard about the ML models that can solve business problems with an ease, all those models rely(depend) on a combination of linear algebra, calculus, probability theory or other math concepts to predict something from some data.
ML models will be trained with structured data
Where should we start?
In the early stages of learning ML ,we used to create models that can give better understanding about the correlations(connections between two or more things) that are being formed using data.
The performance of ML model strongly depend upon
- Data Collection
- Data Pre-processing
- Model Selection
- Training the model
- Evaluating the performance
- Parameter Tuning and
- Make Predictions
Data Collection
It is mainly about collecting data from the sources available. If you are a beginner and you are searching for some best sites to get data-sets then these might help you.
kaggle - Inside you’ll find all the code & data you need to do your data science work. Use over 19,000 public data-sets and 200,000 public notebooks to conquer any analysis in no time.
Data - Here you will find data, tools, and resources to conduct research, develop web and mobile applications, design data visualizations, and more.
Data Pre-processing
It is much needed to get rid of incorrect predictions.
Those include from taking care about missing data to data reduction with same analytical results.
Data Cleaning, Data Integration, Data Transformation, Data Reduction are the steps involved.
Model Selection
This plays a key role in predicting result. A perfect model has the ability to predict results that can transform your business to next level.
Choosing a wrong model leads to incorrect predictions even the business growth is developed but it cannot match the actual results. Mostly the regression models and classification models are used.
Before training the model, one should split the data-set into train set and test set with certain ratio. The most preferable split is 80 -20 that is, 80% of data is for training the model and 20% is to test the model.
Training a Model
The process of training a ML model involves providing an ML algorithm (that is, the learning algorithm) with the training data to learn from.
Before actually predicting result from the model we must train it with lots of data and test it with the data from the test set.
Evaluating the performance
After training the model we should test its performance by predicting the test set results. When we compare the predicted test set results and actual test set the results the performance can be measured.
Parameter Tuning
Tuning certain parameters can increase the accuracy of the predictions by certain models. Parameter Tuning is necessary for certain predictions that deals with more investment.
And Finally, …. Make Predictions.
What are the Libraries used in ML?
There are certain libraries that helps in performing complex operations within a single line of code. By using these libraries, the ML will be more interesting and exiting to explore.
Of Course, the real in-depth coding experience can not be accomplished, but these libraries can save lot of time during the process.
1. NumPy — for large multi-dimensional array and matrix processing.
2. Pandas — for data analysis, data extraction and preparation.
3. SciPy — contains modules for optimization, linear algebra, integration and statistics.
4. Scikit-learn — supports most of the supervised and unsupervised learning algorithms.
5. Theano -is used to define, evaluate and optimize mathematical expressions involving multi-dimensional arrays in an efficient manner.
6. TensorFlow — a framework that involves defining and running computations involving tensors
7. Matpoltlib — is for data visualization.
Even these are some how difficult to understand we will discuss about them in upcoming articles with coding examples.
How is deep learning related to ML?
Deep Learning is a subset of ML, in fact, it’s simply a technique for realizing machine learning.
In other words, Deep Learning is the next evolution of machine learning.
Deep Learning algorithms are roughly inspired by the information processing patterns found in the human brain.
For example, while DL can automatically discover the features to be used for classification, ML requires these features to be provided manually.
Thank you for reading this. I really appreciate all the effort. If you’re having any doubts comment below and I’ll try to answer them.
