MACHINE LEARNING

From: Atul Ranjan Jha

TOPICS
 Introduction to Machine Learning
 Origin of Machine Learning
 Types of Machine Learning
 Machine Learning Principle (work-frame)
 Issue faced by Machine Learning
 Importance of Machine Learning
 Conclusion
 3

INTRODUCTION TO MACHINE LEARNING

Machine learning is a subfield of artificial intelligence (AI) that focuses on
developing algorithms and models that allow computers to learn and make
predictions or decisions without being explicitly programmed for each specific
task. Instead, machine learning algorithms are trained on large amounts of data
and use statistical techniques to recognize patterns, extract insights, and make
predictions or take actions based on the learned patterns.
 ORIGIN OF MACHINE LEARNING
The term Machine Learning was coined by the one of the IBM Worker named Arthur Samuel
in (1952). He wrote a program that played checkers with him and each time it played checkers
it has the tendency to improve and learn from the game it played.
But before Arthur Samuel in (1943) the world’s first mathematical and neural network model
of the Machine Learning was proposed by a mathematician and a neuroscientist named Walter
Pitts and Warren McCulloch and they were able to create algorithms that were able to imitate
the thought process of the humans.
And in (1950) Alan Turing also gave a Turing test which suggested that if a machine is able to
make a human being believe that it is also a human then the machine can be called intelligent.
 6

TYPES OF MACHINE LEARNING

SUPERVISED UNSUPERVISED SEMI-SUPERVISED REINFORCEMENT

Supervised ML learns from Unsupervised ML Semi-supervised ML Reinforcement learning:

labeled data to make discovers hidden patterns leverages both labeled and Agent learns by interacting
predictions based on and structures in data unlabeled data for with the environment to
patterns. without relying on labeled improved learning and maximize rewards.
examples. prediction.
 7

PRINCIPLE OF MACHINE LEARNING

The steps machine learning takes to solve real world problem is:
 Defining the problem: checking the compatibility of the problem if it can be solved by ML and identification
of the data
 Collection of the Data: Collect all the data that are required to solve the problem.
 Data Separation: Separates data that is collected in Training-dataset, Validation-dataset and Test-dataset.
 Algorithm: Choose the Algorithm that can be suitable for the problem that you are solving. ex: Neural
network, etc.
 Training the Model: The above training-datasets are used so that it could train the training model. The training
model are required to learn pattern between input and its corresponding output. Its goal is to minimize the
difference among the value that are predicted by the model and the actual value provided.
 Validation of model: Validate set is used to check validation of model. By its use you can adjust the models
setting.
 Evaluation of the model: Using test set evaluate the performance of the model.
 Deployment and monitor: if all the above step go right then apply this to solve the real-world problem and
continuously monitor the performance.
Atul-jha.
8

ISSUES FACED BY MACHINE LEARNING

INSUFFICIENT TRAINING DATA QUALITY DATA BIASED

DATA
Providing insufficient amount of Noisy/Incomplete data can lead to When a certain data of dataset has
data which is not adequate for less accuracy in the classification of more weight than other and require
training of the algorithm and the data due to which the quality more importance than other it is
contain mostly noisy/incorrect data of the data is low. called data biasing and it leads to
which affect the training of the many errors, damaged outcome,
algorithm. irregular result.
 9

IMPORTANCE OF MACHINE LEARNING

Machine learning has great importance in our life because it has the potential to revolutionize various fields and industries. Some of
the importance are noted down:

1. Data analytics and decision making: ML Algorithms can analyse massive the value of the data and uncover hidden patterns-
progressions, trending things, and insights that may or may not be identified by the humans. This ability to extract valuable
information from data enables better decision making and informed strategies that includes various fields that includes finance,
healthcare, marketing, and logistics.
2. Personalization and recommendation systems: machine learning can help with personalized recommendations, tailored
experiences sectors like E-commerce(online commerce with websites), streaming services/subscription, and
social-media/platform. By analysing user preferences and behaviour, machine learning algorithms that can provide
recommendations for personalized product, suggestions for your content, and target-ed advertising to increase user satisfaction
and engagement.
3. Fraud detection and cybersecurity: These algorithms can also detect pattern/progressions and anomalies in large datasets,
making them valuable for fraud detection and cybersecurity. By analysing historical data and identifying anomalous patterns,
machine learning models can help detect fraudulent transactions, detect network intrusions, and protect against cyber threats.
4. Healthcare and medical sector : Machine learning has the potential to revolutionize healthcare by enabling more accurate
medical diagnoses.
 10

SUMMARY
• In conclusion, this machine learning project aimed to address a
specific problem or task using various algorithms and techniques.
Through data preprocessing, feature engineering, model selection,
and evaluation, we have successfully developed a machine
learning model that provides valuable insights and predictive
capabilities.

• Machine learning faces various issues and challenges and also

provides various benefit in various sector like: automation,
medical, finance, etc. Despite the challenges, machine learning
continues to evolve, offering immense potential to solve complex
problems, drive innovation, and improve decision-making in
diverse fields. By recognizing the benefits and addressing the
issues, we can harness the power of machine learning responsibly
and ethically to create a positive impact on society.
"Machine Learning, where data empowers intelligence's gleam“
– Unknown

THANK YOU..