Get started with this applied course focusing on the introduction to data preprocessing. This quick course provides opportunities of learning through reading material, Business case and a quiz to test your understanding. Beginner Level Approx. 1 Hour to complete
This course will help you understand the need of explainable AI (XAI) and introduce you to LIME, which is one of techniques for XAI. Learn how LIME can be applied on a real world datacase and how the results can be presented in a human interpretable format. Take the quiz to validate your understanding and then solve a datacase based on the learnings.
This course helps you to understand different types of basics of summary and distribution charts which are majorly used to represent the data in a visible format to reveal the trends and patterns. Learn how and when to use different types of charts to present the actionable insights.
Analysis of Variance (ANOVA) is a statistical method to test difference between two or more means. This applied course focuses on the concepts of ANOVA , its types and applications by using business cases. Basic Level
This course will help you understand the need of explainable AI (XAI) and introduce you to SHAP, which is one of techniques for XAI. Learn how SHAP can be applied on a real world datacase and how the results can be presented in a human interpretable format. Take the quiz to validate your understanding and then solve a datacase based on the learnings.
Gradient Boosting Classification is a machine learning technique, attracting attention for its prediction and accuracy. From Machine Learning competitions to data science businesses gradient boosting is providing best-in-class results. It uses an ensemble of weak prediction models for building a classification model. The model is built in a sequential process, where each new step tries to minimize the errors of the previous step. This course will provide you an in-depth understanding of the algorithm along with its practical use case.
Extreme Gradient Boosting is an optimized machine learning algorithm designed for better speed and performance. It is an extension of gradient boosted decision trees. Since its inception in 2014, the algorithm has achieved high performances on structured data and dominated other algorithms because of its flexibility. This course covers the end-to-end walkthrough of the algorithm.
CatBoost (short for Categorical Boosting) is a recently developed machine learning algorithm that helps data scientists and machine learning engineers in automating their column transformations and provides best-in-class accuracy. It can easily integrate with different deep learning frameworks and can work on diverse data types. It solves a variety of business problems that earlier required extensive data preparation.
This course will give you a brief introduction to the concept and applications of one of the most important and commonly used model validation techniqes: Gain and Lift Charts.
One of the most essential phases of the classification machine learning algorithms is to understand the ways to evaluate the performance of the model built. This is one of the tasks that are often difficult to interpret and sometimes not paid attention to. In this course, we will see each of the performance measures that will help you to quantify the performance of the model and decide which algorithm will be most suited according to the problem at hand.
Machine Learning algorithms have been proven extremely beneficial for regression tasks. However, there is one question that is often asked. Which algorithm is suited for the particular problem or Do you think the model built is performing well? In this course, we will learn different performance measures that will help you to answer this question. We will see each performance measure used to evaluate a regression model in detail which will help us to select the top-performing model for a particular problem.
When we picture Artificial Intelligence, we often picture ourselves talking to a machine. In fact, we interact with personal assistants (Siri, Cortana, Alexa) on a daily basis and give them instructions to perform tasks. This functionality has been enabled through the domain of NLP. NLP is concerned with the interactions between computer and human languages. Machine Learning algorithms are applied to text and speech data in NLP.
Raw or source data is often inconsistent, vague, and duplicative in nature. Therefore, Data transformation is the process that helps convert this raw structure to another structure that can help you summarize and discover knowledge. The Data Transformation process is essential to any business, especially when there is a need to maintain data quality to deliver recommended Business Intelligence and a requirement to change the structure to a more readable format, enabling effective data analysis.
Data grows exponentially every day and it is becoming an essential factor for driving businesses. Huge amounts of data are being generated and from different sources. Hence, it is essential to perform data aggregation to identify patterns and trends in the data which otherwise would not have been recognized in the data from a single source. This course will enable you to understand different data aggregation methods that are being used on a regular basis in any data-based decision-making.
Feature Engineering is the process of using domain knowledge to identify or create the important features that have the strongest predictive power. It is considered one of the fundamental tasks for improving machine learning model performance and prediction accuracy. In this course, you’ll learn a number of techniques for selecting the most important features for model prediction.
The course on ‘Model Selection Techniques’ introduces the concept of various techniques used for selecting better predictive models based on the different number of available models. It helps in differentiating the models based on different criterions available in the data science theory. Model selection uses important principles of data analytics and helps in selecting better model, thereby improving predictability of the model.
Overfitting is a common phenomenon in real-world data. Regularization is the technique that helps deal with overfitting problem in machine learning. Regularization is used to put additional constraints on the loss function of the model. It introduces bias in the model by reducing complexity but the variance is lowered as well. In this course, the concepts of regularization, Bias-variance tradeoff are discussed explaining the fit of the machine learning model.
Bagging is a technique in which predictions of multiple machine learning models are aggregated to make the final prediction by either majority rule or aggregation function. One of the most useful algorithms based on the bagging ensemble technique is Random Forest. Through this course, let us understand the components of bagging and random forest and apply ensemble algorithms to find solutions to real-world problems.