|44.50||ca. 7-9 Tage|
Microsoft Excel MVP Conrad Carlberg shows readers how to use Excel predictive analytics to solve real-world problems in areas ranging from sales and marketing to operations. Carlberg offers unprecedented insight into building powerful, credible, and reliable forecasts, showing how to gain deep insights from Excel that would be difficult to uncover with costly tools such as SAS or SPSS.
Readers will get an extensive collection of downloadable Excel workbooks that can be easily adapted to their own unique requirements, plus VBA code—much of it open-source—to streamline several of this book’s most complex techniques.
Step by step, readers build on Excel skills they already have, learning advanced techniques that can help them increase revenue, reduce costs, and improve productivity.
Chapter 1 in the first edition of the proposed book concerned itself with how to use
Excel and VBA to download periodically (e.g., by second, by minute, by hour, by day, etc.) sales and related information from web sites. This information did not get much comment in Amazon reviews (there are 42 reviews as of November 2016) and what comment there was split itself between those who like to know how to do that sort of data acquisition, and those whose background in VBA is far too sketchy to understand what was going on. In the interim, many such sites have instituted bot blockers that refuse to complete a query not sent by a browser. I think that it makes sense to drop this chapter but could be talked into retaining it.
The existing Chapter 4 contains information on two types of exponential smoothing: simple smoothing and Holt's linear smoothing, which takes account of trend in a time series. I would expect to expand the chapter to provide more information on Holt's method and add a discussion of Holt-Winters smoothing, which accounts for seasonality in a time series. I might add some material on damped trend forecasts, which help prevent trends from getting out of control beyond the normal forecast horizon.
I also intend to add material to the existing chapter 7 on more advanced issues in logistic regression. At present, the discussion covers situations in which the outcome variable has two values only. Logistic regression also deals with situations in which the outcome variable has 3 or more categorical values. That's a considerably more complex situation. It's difficult to illustrate in Excel and the R syntax to deal with it is esoteric. But the situation arises frequently, particularly in consumer choice situations, and it should get coverage in this book.
Chapter 1 Building a Collector
Chapter 2 Linear Regression
Chapter 3 Forecasting with Moving Averages
Chapter 4 Forecasting a Time Series: Smoothing
Chapter 5 Forecasting a Time Series: Regression
Chapter 6 Logistic Regression: The Basics
Chapter 7 Logistic Regression: Further Issues
Chapter 8 Principal Components Analysis
Chapter 9 Box-Jenkins ARIMA Models
Chapter 10 Varimax Factor Rotation in Excel