Twitter LinkedIn Facebook RSS Android

Data Mining and Predictive Analytics Software For Microsoft Excel: Page 2

By Vangie Beal     Feedback
Previous |   Page 2 of 3   |   1 2 3   |   Next /images/next-horizontal.png

Data Mining and Predictive Analytics in Four Steps

The tool allows you to mine your own data to find patterns using only four steps — prepare, analyze, predict and report. It is also designed to automate the process of algorithm selection, parameter tuning and reporting. Each step is easily accessed using a separate tab in your 11Ants Model Builder Excel Ribbon.

Prepare: By selecting data columns in your spreadsheet, you choose one column as the target column and others as the input. For example, for sales data you might use season, date and volume as the input and the revenues column as the target. The 11Ants Model Builder will analyze the relationships between the input and the target. The data is then partitioned — or split — into two sets: training and test. Here users familiar with Excel will have little problem with getting the data prepared.

Options in Model Builder allow you to change the target column and adjust the weight between test and train size. Once you have selected options, you select "Prepare Sheets" and you will find your Excel worksheet is now three worksheets: the original plus one for training data and test data.

Analyze Data and Generate Models: In the training data sheet you, can analyze the data by assigning a data type to a column (category, number or date/time) and then choosing the type of prediction. The 11Ants Model Builder offers three types of prediction: predict number, predict category, or rank likelihood. After choosing options you continue by selecting "Start Generating Models."

As this process runs, 11Ants Model Builder is analyzing the data for relationships and generates continuous models looking for the best one. The quality score changes based on the amount and quality of data being analyzed.

You can view quick info about the project, including estimated Input influences, Top 10 and improvement curve. As the process runs, you can watch the quality score — the higher the percentage, the more patterns found in the data. This information is also available through the Manage Tab in the Excel Ribbon.

Predict: When ready, you can build your predictive model using the test spreadsheet. By choosing "Predict" from the ribbon, you can choose your prediction settings and a new sheet will be generated to see how the model works on the test data. Using your test data worksheet, you can choose your model, confirm your input data type, choose a column to output the results, assign a confidence (hi/med/low) to each prediction, and also decide the type of prediction to be reported and compare predictions against known values.

When you click "Predict Now," a new worksheet for the prediction statistics is generated.


This article was originally published on December 17, 2010
Previous |   Page 2 of 3   |   1 2 3   |   Next /images/next-horizontal.png
Close Icon
Thanks for your registration, follow us on our social networks to keep up-to-date