What is Machine Learning?

by Dec 1, 2021#MachineLearning, #HomePage

Printer Icon

Machine Learning is an application in which machines can learn from their experiences or train data to make predictions. The Machine Learning approach is different from traditional programming in that the computer learns automatically, detecting patterns and creating its own rules, thereby making it more accurate and easier to maintain.

According to Stanford University professor Andrew Ng, “Machine Learning (ML) is the science of getting computers to act without being explicitly programmed.” Instead of writing code, the user feeds the dataset to the generic algorithm, and the algorithm or the machine will operate with logic based on the given data. Just as our brains use our experience to help us improve at a task, so does the computer.

For example, let’s say you want a computer that can tell the difference between a picture of a car and a picture of a bus. So, you begin updating images that show cars and buses. The computer has to figure out that cars are smaller in size and have several variants, whereas buses are different and larger.

Then, in the future, when the computer sees a picture, it will check the picture’s pattern and decipher if it is a car or a bus. There can (and will) be mistakes, of course, but the algorithm will become more accurate in its predictions over time as it receives more data.

The Machine Learning model is becoming more and more prevalent in daily life in the 21st century. Considered one of the most significant innovations since the microchip, ML has the potential to transform our world in truly mind-blowing ways. Consider how Machine Learning has already impacted our daily lives:


Applications that Use Machine Learning

  • Speech recognition—Alexa, Google Assistant, Google Home, Siri
  • Face or image recognition (automatic friend-tagging suggestions)
  • Self-driving cars by Tesla
  • Recommendations on Netflix, Spotify, YouTube, etc.
  • Google Search
  • Google Maps (showing traffic conditions)
  • Spam filters on emails
  • Medical diagnoses and healthcare
  • Online fraud detection
  • Online shopping with Amazon, eBay, etc.
  • Intelligent Gaming—AlphaGo, Deep Blue, etc.
  • Social networking—Facebook, Instagram, Twitter, etc.

Types of Machine Learning Systems

Machine Learning systems are classified according to how they are trained to learn incrementally, how they generalize, and how data points are compared or built to detect patterns. The output model can be combined with other Machine Learning systems and data components to create the right solution.


Supervised Machine Learning:

Supervised learning, as the name indicates, involves a supervisor’s presence as a teacher and a training set that includes the solutions (labels). This is the simplest form of Machine Learning, a method where you have input variables (X) and output variables (Y), which implies Y = f(X). Our end goal is to approximate the mapping function (f) so that we can predict the output variables (Y) when we have new input data (X). This training dataset includes inputs and correct outputs, which allows the model to learn over time. For example, a machine learns to classify whether an image is of a bird or an animal.


Unsupervised Machine Learning:

As the name indicates, in the case of unsupervised learning, there is no help from the user for the computer to learn, i.e., no labeled training sets. This allows the model to work on its own to discover patterns and information previously undetected. There are no actual data points in unsupervised learning, and references are drawn from observations in the input data. For example, an ML model could help a user understand different client groups around which to build a business strategy.


Semi-Supervised Machine Learning:

Semi-supervised machine learning combines supervised and unsupervised learning by using a few labeled data and plenty of unlabeled data, which helps avoid the challenge of finding large amounts of labeled data. This model is trained to label data.

Categories of Semi-Supervised Learning Methods:

  • Inductive Learning (Inference): In this method, the model learns from a specific dataset and generalizes to make predictions on unseen data.
  • Transductive Learning: This method refers to reasoning from specific observed (training) instances to specific observed (unlabeled) instances. An example would be a text document classifier. A semi-supervised learning algorithm can label data and retrain the model with the newly labeled dataset.

Reinforcement Learning:

In reinforcement learning, the model learns with a rewarding process (positive or negative) to perform tasks in a better way, based on the actions or experiences

Primary Components of Reinforcement Learning:

  • Agent (the learning system)
  • Environment (agent comes in contact with the environment to select actions)
  • Reward or penalty
  • Policy (learn the strategy—define the appropriate action in a given situation)
  • Iterate (the update policy)

An agent learns from the environment by interacting with it—taking the necessary action to achieve the best result—and learns to create a “policy” that defines future actions.


Batch Learning/Offline Learning:

In batch learning, the ML system is trained using the total data available. The model only works with a limited set of data and does not learn incrementally. When new data becomes available, the system is updated to a new version. Batch learning algorithms are used for small quantities of data with no incoming data.

Online Learning/Incremental Learning:

In online learning, the ML system learns incrementally with sequential instances. This method is best for systems with continuous flow and fast changes in the data.


Instance-Based Learning:

In batch learning, the ML system is trained using the total data available. The model only works with a limited set of data and does not learn incrementally. When new data becomes available, the system is updated to a new version. Batch learning algorithms are used for small quantities of data with no incoming data.

Model-Based Learning:

In online learning, the ML system learns incrementally with sequential instances. This method is best for systems with continuous flow and fast changes in the data.

Techniques to Implement Machine Learning

Linear Regression:

The regression method belongs to the category of supervised ML. Linear regression is the simplest and most common method of learning predictive modeling. It is used to estimate real values (cost of houses, number of calls, total sales, the stock price will increase or decrease, etc.) based on continuous variables. In linear regression, the relationship between the input variables (x) and output variable (y) is expressed as an equation: Y = a + bX. Thus, linear regression aims to determine the values of coefficients and tries to fit data with the best hyperplane that goes through the points.


Classification is an essential component for AI applications and is also required for e-commerce applications. This method allows us to make more informed decisions—sort out spam, predict whether a borrower will return a loan, predict whether or not an online customer will buy a product, conduct fraud detection, tag friends in a Facebook image, and so on. These algorithms predict discrete variable labels. Several classification models are logistic regression, decision tree, random forest, gradient-boosted tree, multilayer perceptron, one-vs-rest, and Naive Bayes.


Clustering algorithms are unsupervised learning methods which group the unlabeled dataset. This method develops collections of objects based on similarity and dissimilarity. Clustering is widely used in sales and marketing for customer segmentation and personalized communication. For example, Amazon and Netflix use clustering to provide new recommendations based on a past search. A few common clustering algorithms are k-means clustering, mean-shift, and expectation-maximization.

Decision Tree:

This is a supervised learning algorithm mainly used for classification problems and regression problems. A decision tree asks a question and, based on the answer (Yes/No), it further splits the tree into subtrees. A typical example of a decision tree would be identifying the insurance premium that someone should be charged based on that individual’s situation.


Machine Learning has become a key component of business operations and digital strategy, together with business data and computing power has been transforming businesses and ecosystems.

But powering your products with Machine Learning algorithms requires high-level skills and a transformational culture.

Products can mine data from Machine Learning algorithms and find patterns in data to build models that make predictions and scale products to a more sophisticated level.

Machine Learning puts users at the center of the business to solve a problem, builds a business case, and applies data science and engineering.

Want to learn more about how to empower your business products or services?  how to discover the right algorithms?  or compare ML models and simulations?

About Us: Krasamo is a mobile-first Machine Learning and consulting company focused on the Internet-of-Things and Digital Transformation.

Click here to learn more about our machine learning services.


Feature Engineering for Machine Learning

Feature Engineering for Machine Learning

Feature engineering is a crucial aspect when it comes to designing machine learning models, and it plays a big role in creating top-notch AI systems. Features are attributes that represent the problem of the machine learning use case and contribute to the model’s prediction. The process of feature engineering involves creating relevant and useful features from raw data combined with existing features, adding more variables and signals to improve the model’s accuracy and performance. It starts manually and can be accelerated by adding automated feature engineering tools and techniques. Follow the steps of feature engineering to optimize your machine learning models and create innovative products.

Machine Learning in IoT: Advancements and Applications

Machine Learning in IoT: Advancements and Applications

The Internet of Things (IoT) is rapidly changing various industries by improving processes and products. With the growth of IoT devices and data transmissions, enterprises are facing challenges in managing, monitoring, and securing devices. Machine learning (ML) can help generate intelligence by working with large datasets from IoT devices. ML can create accurate models that analyze and interpret the data generated by IoT devices, identify and secure devices, detect abnormal behavior, and prevent threats. ML can also authenticate devices and improve user experiences. Other IoT applications benefiting from ML include predictive maintenance, smart homes, supply chain, and energy optimization. Building ML features on IoT edge devices is possible with TensorFlow Lite.

DataOps: Cutting-Edge Analytics for AI Solutions

DataOps: Cutting-Edge Analytics for AI Solutions

DataOps is an essential practice for organizations that seek to implement AI solutions and create competitive advantages. It involves communication, integration, and automation of data operations processes to deliver high-quality data analytics for decision-making and market insights. The pipeline process, version control of source code, environment isolation, replicable procedures, and data testing are critical components of DataOps. Using the right tools and methodologies, such as Apache Airflow Orchestration, GIT, Jenkins, and programmable platforms like Google Cloud Big Query and AWS, businesses can streamline data engineering tasks and create value from their data. Krasamo’s DataOps team can help operationalize data for your organization.

What Is MLOps?

What Is MLOps?

MLOps are the capabilities, culture, and practices (similar to DevOps) where Machine Learning development and operations teams work together across its lifecycle

ETL Pipelines and Data Strategy Overview

ETL Pipelines and Data Strategy Overview

Data is a primary component in innovation and the transformation of today’s enterprises. But developing an appropriate data strategy is not an easy task, as modernizing and optimizing data architectures requires highly skilled teams.