DATASCIENCE ONLINE TRAINING
Basic Conceptsof Statistics:
1. Descriptive Statistics and Probability
Distributions:
·
Introduction about Statistics
·
Different Types of Variables
·
Measures of Central Tendency with examples
§
Mean
§
Mode
§
Median
·
Measures of Dispersion
§
Range
§
Variance
§
Standard Deviation
·
Probability & Distributions
·
Probability Basics
·
Binomial Distribution and its properties
·
Poisson distribution and its properties
·
Normal distribution and its properties
2. Inferential Statistics and Testing of Hypothesis
·
Sampling methods
§
Sampling and types of sampling
§
Definitions of Sample and Population
§
Importance of sampling in real time
§
Different methods of sampling
§
Simple Random Sampling with replacement and without replacement
§
Stratified Random Sampling
· Different methods of estimation
·
Testing of Hypothesis & Tests
§
Null Hypothesis and Alternate Hypothesis
§
Level of Significance and P value
§
t-test and its properties
§
Chi-square test and it’s properties
§
Z test
· Analysis of Variance
§
F-test
§
One and Two way ANOVA
3. Covariance & Correlation
§
Importance and Properties of Correlation
§
Types of Correlation with examples
Predictive Modeling Steps and Methodology with Live example:
· Data Preparation
§
Variable Selection
§
Transformation of the variables
§
Normalization of the variables
·
Exploratory Data analysis
§
Summary Statistics
§
Understanding the patterns of the data at single and multiple dimensions
§
Missing data treatment using different
methods
§
Outlier’s identification and treating
outliers
§ Visualization
of the data using the One Dimensional, Two Dimensional and Multi Dimensional Graphs.
Bar chart,
Histogram, Box plot, Scatter plot, Bubble chart, Word cloud etc…
·
Model Development
§
Selection of the sample data
§ Selecting
the appropriate model based on the requirement and data availability
·
Model Validation
§
Model Implementation
§
Key Statistical parameters checking
§
Validating the model results with the actual result
·
Model Implementation
§
Implementing the model for future prediction
· Real time telecom business
use case with detail explanation
· Introducing couple
of real time use cases and solutions
of Banking and Retail
domains using the different statistical methods.
Supervised Techniques:
·
Multiple linear Regression
§
Linear Regression - Introduction - Applications
§
Assumptions of Linear Regression
§
Building Linear Regression Model
§ Understanding
standard metrics (Variable significance, R-square/Adjusted
R-Square, Global hypothesis etc)
§ Validation of Linear
Regression Models (Re running Vs. Scoring)
§ Standard
Business Outputs (Decile Analysis, Error distribution (histogram), Model equation, drivers etc)
§
Interpretation of Results - Business Validation - Implementation on new data
§ Real time
case study of Manufacturing and Telecom Industry to estimate the future revenue using the models
· Logistic Regression
§
Logistic Regression - Introduction -
Applications
§
Linear Regression Vs. Logistic Regression Vs. Generalized Linear Models
§
Building Logistic Regression Model
§ Understanding
standard model metrics (Concordance, Variable
significance, Hosmer Lemeshov Test, Gini, KS, Misclassification etc)
§ Validation of Logistic
Regression Models (Re running Vs. Scoring)
§
Standard Business Outputs (Decile Analysis, ROC Curve)
§
Probability Cut-offs, Lift charts, Model equation, drivers etc)
§
Interpretation of Results - Business Validation - Implementation on new data
§ Real time
case study to Predict the Churn customers in the Banking and Retail industry
· Partial Least Square Regression
§
Partial Least square Regression - Introduction - Applications
§
Difference between Linear Regression and Partial Least Square Regression
§
Building PLS Model
§ Understanding
standard metrics (Variable significance, R-square/Adjusted
R-Square, Global hypothesis etc)
§
Interpretation of Results - Business Validation - Implementation on new data
§ Sharing the
real time example to identify the key factors which are driving the Revenue
Variable Reduction Techniques
·
Factor Analysis
·
Principle component analysis
§
Assumptions of PCA
§
Working Mechanism of PCA
§
Types of Rotations
§
Standardization
§
Positives and Negatives of PCA
Supervised Techniques Classification:
·
CHAID
·
CART
·
Difference between CHAID and CART
·
Random Forest
§
Decision tree vs. Random Forest
§
Data Preparation
§
Missing data imputation
§
Outlier detection
§
Handling imbalance data
§
Random Record selection
§
Random Forest R parameters
§
Random Variable selection
§
Optimal number of variables selection
§
Calculating Out Of Bag (OOB) error rate
§
Calculating Out of Bag Predictions
·
Couple of Real time use cases which are related to
Telecom and Retail Industry.
Identification of the Churn.
Unsupervised Techniques:
·
Segmentation for Marketing Analysis
§
Need for segmentation
§
Criterion of segmentation
§
Types of distances
§
Clustering algorithms
§
Hierarchical clustering
§
K-means clustering
§
Deciding number of clusters
§
Case study
·
Business Rules Criteria
·
Real time use case to identify the Most Valuable revenue generating Customers.
Time series Analysis:
·
Forecasting - Introduction - Applications
·
Time Series Components( Trend, Seasonality, Cyclicity
and Level) and Decomposition
· Basic Techniques –
§ Averages,
§ Smoothening etc
· Advanced Techniques
§ AR Models,
§ ARIMA
§ UCM
§ Hybrid Model
·
Understanding Forecasting Accuracy - MAPE, MAD, MSE etc
·
Couple of use cases, To forecast the future sales of products
Text Analytics:
·
Gathering text data from web and other
sources
·
Processing raw web data
·
Collecting twitter data with Twitter
API
· Naive Bayes Algorithm
§ Assumptions and of Naïve Bayes
§ Processing of Text data
§ Handling Standard and Text
data
§ Building Naïve Bayes Model
§ Understanding standard model
metrics
§ Validation of the Models (Re
running Vs. Scoring)
· Sentiment analysis
§ Goal Setting
§ Text Preprocessing
§ Parsing the content
§ Text refinement
§ Analysis and Scoring
·
Use case of Health care industry, To identify the
sentiment of the patients on
Specified hospital by extracting the data from the TWITTER.
Visualization Using Tableau:
·
Live connectivity from R to Tableau
·
Generating the Reports and Charts
R PROGRAMMING
SESSION 1: Getting Started
with R
· What is statistical programming?
· The R package
· Installation of R
· The R command line
· Function calls, symbols, and assignment
· Packages
· Getting help on R
· Basic features of R
·
Calculating
with R
SESSION 2: Matrices, Array, Lists,
and Data Frames
·
Character vectors
·
Operations
on the logical vectors
·
Creating
the matrices and operations on it
·
Creating
the array and operations on it
·
Creating
the lists and operations on it
·
Making
data frames
·
Working
with data frames
SESSION3: Getting Data in and out
of R
·
Importing
different types of file formats
SESSION4: Data Manipulation and
Exploration:
· Variable transformations
· Creating Dummy variables
· Data set options (Rename, Label)
· Identification and Dealing with the
Missing data
· Sorting the data
· Handling the Duplicates
· Summarize numeric variables
· Summarize factor variables
· Aggregated functions using Group by
·
Data
preparation using the sqldf package
SESSION5: Conditional Statements
and Loops:
· Nested If Else
· For Loop
·
While Loop
SESSION6: Functions:
· Numeric Functions
SESSION7: Graphical procedures
· Bar Chart
· Box plot
· Scatter plot
· Multi Scatter plot
·
Word cloud etc.…
SESSION8:
Advanced R and Real time analytics examples:
· Text Data handling
· Positive and Negative word cloud
· Required packages for the analytics
· Sentiment analysis using the real
time example
· R code automation
· Time series analysis with the real
time Telecom data
· Couple of examples with the time
series data
