Data Science-Forecasting/Time series Using XLMiner,R&Tableau
- 6.5 hours on-demand video
- 1 article
- 4 downloadable resources
- Full lifetime access
- Access on mobile and TV
- Certificate of Completion
Get your team access to 4,000+ top Udemy courses anytime, anywhere.Try Udemy for Business
- Learn about different types of approaches using XLminer, R and Tableau
- Learn about the Forecasting Importance ,Forecasting Strategy which includes Defining goal, Data Collection, Exploratory Data Analysis, Partition Series, Pre-process Data, Forecast Methods, using various Plots.
- Learn about scatter diagram, correlation coefficient, confidence interval, which are all required for implementing forecasting techniques
- Learn about the various error measures such as ME, MAD, MSE, RMSE, MPE, MAPE, MASE
- Learn about Model based Forecasting Techniques such as Linear, Exponential, Quadratic, Additive Seasonality, Multiplicative Seasonality, etc.
- Learn about Auto Regressive Models for using errors to further strengthen the forecasting model used & also learn about Random walk & how to identify the same
- Learn about Data Driven approaches such as Moving Average, Simple Exponential Smoothing, Double Exponential Smoothing / Holts, Winters / HoltWinters
Learn about the trainer introduction & course introduction, which briefly explains about the concepts to be discussed through out the program.
Walk through the steps in performing Additive seasonality with trend forecast and multiplicative seasonality; Configuring visually the best model and ratifying with the least error model; Combining the training and Validation data to run the final model
Packages in R for reading file; creating a file access pop up using R script; create a data frame with dummy variable for seasonality; creating & assigning new variable to a data frame as a part of pre-processing; evaluating various prediction models for forecasting namely Linear Model, Exponential Model, Quadratic Model (Polynomial trend with two degrees); performing the predict function on the test data for each of the models and evaluating the RMSE(Root Mean Square Error)
Various other model based prediction strategies viz, Additive seasonality, Additive seasonality with Quadratic trend, Multiplicative seasonality; and evaluating for the model with the least RMSE; Running the model with least error using the complete data base for making the prediction; performing an ACF plot on the residuals from the model; Auto regressive model AR(1) on the residuals; check for significant information left over in the residuals of residuals
Features of autoregressive models, Detailed study of residuals and its leftover information, Forecasting errors by AR model using the principle of Parcimony, iterative study of ‘residuals of residuals’ till no further information can be gleaned, Building the final model by plugging in the error factor
Computing the Moving average even window width and walk thru of the steps in computing the seasonal indexes, Normalizing the seasonal indexes and drawing inferences for these, computing the de seasoned data using the respective seasonal indexes
Performing simple moving average using ARIMA; Selecting the appropriate MA value based on the Seasonality; performing the forecast() function and the accuracy function to check errors for training and test data; Performing simple exponential smoothing using only the level component of the time series data; Building Double exponential smoothing model using the level and trend component of the time series data; performing Holts winters model using level, trend and seasonality;
Performing the various exponential smoothing models using the HoltsWinters() command, so as to determine the optimum alpha, beta and gamma values corresponding the Level, Trend and seasonality; Performing the various smoothing techniques using commands line ses() for Simple exponential smoothing, holt() for Holts and hw() for holt Winters technique; Determining the best model based on the MAPE (Mean absolute percent error); Running the best model to from the list on the total data and forecasting one full cycle of seasonality
Forecasting in Tableau using exponential smoothing; choosing date dimension in the column shelf and value dimension in the rows self; Selecting the analytics pane to choose the forecast option for obtaining the forecasts; Understanding the various options within the forecast; Viewing the errors of the forecast; forecasting for non time series numerical data; The eight types of forecasts models;
- Download XLminer, R , RStudio before starting this tutorial
- Download datasets folder in zip file which is uploaded in Session 1
Forecasting using XLminar,Tableau,R is designed to cover majority of the capabilities from Analytics & Data Science perspective, which includes the following
- Learn about scatter diagram, autocorrelation function, confidence interval, which are all required for understanding forecasting models
- Learn about the usage of XLminar,R,Tableau for building Forecasting models
- Learn about the science behind forecasting,forecasting strategy & accomplish the same using XLminar,R
- Learn about Forecasting models including AR, MA, ES, ARMA, ARIMA, etc., and how to accomplish the same using best tools
- Learn about Logistic Regression & how to accomplish the same using XLminar
- Learn about Forecasting Techniques-Linear,Exponential,Quadratic Seasonality models,Linear Regression,Autoregression,Smootings Method,seasonal Indexes,Moving Average etc,...
- All the IT professionals, whose experience ranges from '0' onwards are eligible to take this session. Especially professionals from data analysis, data warehouse, data mining, business intelligence, reporting, data science, etc, will naturally fit in well to take this course.