Data Science Training

As Data Science is a broader concept or multidisciplinary subject, machine learning also falls under Data Science segment. In this section, you will be introduced to various machine learning concepts.  

Topics covered in this section are:

• Introduction to machine learning, linear regression, and predictive modelling 

• Modelling with simple linear, multiple Linear regression and Linear regression 

• Finding P-value, making the comparison between Linear regression and logistic regression, 

• Evaluation with ROCR, detailed formulas, and understanding the fit of a model 

• Predicting results, understanding the summary results with the null hypothesis. 

• Building Linear models with multiple independent variables.             

Learning Outcome: By the end of this module, you will become an expert in using the linear predictor for modelling the relationships within the data. You will also become familiar with various regressions and in implementing them.

This section deals with complete regression concepts, their importance, and how they work in real-time. It contains complete details regarding regression concepts and usage. 

Topics covered in this section are: 

• Introduction to logistic regression, its concepts, and Linear vs Logistic regression  

• Math behind linear regression concept, detailed formulas, and logit and odds

• Building simple “binomial” model, confusion matrix, accuracy, true positive rate, false-positive rate, etc

• Introduction to confusion for evaluating the built model and finding the right threshold by building the ROC plot.

• Concepts like cross-validation and building logistic models based on real-life applications of Regression. 

Learning Outcome: By the end of this session, you will be able to describe data and explains the relationship between one or more binary variables. You will also come to know how to use glm() to build models.

This section is meant to deal with fundamental concepts of Decision Tree & Random Forest which include classification techniques algorithms, creating decision trees and classification trees, and all other core concepts. 

Topics covered in this section are:    

• Introduction to classification and other classification techniques 

• Introduction to Decision tree & algorithms for decision induction, and building decision tree in R. 

• The process to create a perfect decision tree and confusion matrix. 

• Introduction to the ensemble of trees, Random Forest, bagging, implementing Random Forest in R. 

• Computing probabilities, light split node, Information gain, Gini index for a right split of the node.  

• Cost complexity pruning, Pre-pruning, and post-pruning 

• Finding the perfect number of trees and evaluate performance metrics. 

Learning Outcome: 

You will gain expertise in regression & classification concepts. You will be able to build a tree & prune it by taking the help of Churn, and build a Random forest with the right number of trees.

Unsupervised learning occupies an important position in Data Science. It is a type of self-organized Hebbian learning which helps in finding previously unknown patterns in data without the need for having a pre-existing label and allows modelling in probabilities densities of given inputs. 

Topics covered in this section are: 

• Introduction to Unsupervised learning, clustering & its use cases

• Concepts like Canopy clustering, Hierarchical clustering, and Theoretical aspect of K-means  

• Introduction to Unsupervised learning, Clustering algorithms, K-means in the process flow, K-means ion process Flow, and K-means in R 

• Finding the right number of clusters with the help of Scree-plot Dendrogram & clustering 

• Understand Hierarchical clustering and implementation of R within hierarchical clustering,

• Introduction to Component analysis, PCA in R, and procedure to Implement PCA  

Learning Outcome: You will gain hands-on expertise in unsupervised learning with R, dimensionality reduction and how K-means clustering works for visualizing and interpreting data.

Association rule learning is a method used in machine learning for discovering interesting relations between variables in a large data set. And, the other topic is a recommendation engine, which is a software algorithm used to analyze the available information and to recommend relevant information to the user. 

Topics covered: 

• Introduction to rule mining, the measure of association rule mining, Apriori algorithm and its implementation process in R

• Introduction to a recommendation engine, collaborative user-based filtering, & Item-based collaborative filtering.

• Implementing recommendations in R, recommendation use-cases: user-based and item-based. 

Learning Outcome: By the end of this session, you will be able to deploy association analysis as a rule-based method. You will be able to discover strong rules in the database based on interesting discoveries.

This section deals with the most important topic, artificial intelligence, and how it is associated or related to Data Science. 

Topics covered are: 

• Introduction to Artificial intelligence & Deep learning 

• Introduction to TensorFlow computational process & Artificial Neural Network 

• A process on how to build Artificial Neural Networks using TensorFlow, and working with TensorFlow in R.    

Learning Outcome: By the end of this session, you will gain complete knowledge of concepts like Artificial intelligence TensorFlow and hands-on experience in building ANN Frameworks.

Data science is used these days in all industries, whether it is IT, Healthcare, Insurance, Finance, Entertainment, or Media. 

Association rule and recommendation engine case study

This case study will help you in understanding the association rule, recommendation engine, data analytics tools such as Hadoop, Spark and how to leverage these tools for the business growth. For example, Netflix, the largest internet-television network used big data analytics to understand the users’ needs and by focusing on their requirements, increased the customers base. Its reality TV shows work on the recommendation approach, which also brings in customer’s loyalty. This is a perfect example for the business that demands constant engagement and association of users for its growth and success.

Reverse engineering case study

In this case study, you will learn predictive modelling and reverse engineering. The Mint.com website is one such example, which provides free insights of where money goes to for its investors. It grew from zero to 100, 000 members in just six months. It followed the reverse engineering marketing methodology to gain the faith in the product and started working on the solution to way back to the individual data.

Statistical models case study

This case study focuses on big data analytics and parallels statistical models. Pharmaceutical industries get the advantage of such tools to prioritize the products for the treatment. The parallel statistical model helps in forecasting the likelihood of doctors prescribing medicine and henceforth, demand for that product. You will prepare a real-time model of how these tools can be helpful for clinical research scientists to formulae medicine. 

Predictive analytics and decision tree case study

You will learn how to use predictive analytics and decision tree in this case study. You can use these tools for the education industry to forecast the dropout rates of students based upon their performance. For example, the University of Florida uses IBM Cognos Analytics to focus on the student's performance vs dropout ratio. Accordingly, it provides solutions to retain students.