---

KXEN Event Log (KEL) aggregates events into periods of time. KEL allows integrating transactional data with demographic customer data. It is used in cases when the raw data contains static information such as age, gender or profession of an individual, and dynamic variables, such as spending patterns or credit card transactions. Data is automatically aggregated within user defined periods without programming SQL or changing database schemas. KEL combines and compresses this data to make it available to other KXEN components.

Benefits: KEL allows you to integrate additional sources of information on the fly to improve your model quality.

KXEN Sequence Coder (KSC) aggregates events into a series of transitions. For example a customer click-stream from a Web site can be transformed into a series of data for each session. Each column represents a specific transition from one page to another. Similar to KEL these new columns of data can be added to existing customer data and are made available to other KXEN components for further processing.

Benefits: With KSC you can tap into previously unused sources of information to build better predictive models.

KXEN Consistent Coder (K2C) automatically prepares and transforms data into a format suitable for use in the KXEN Analytic Framework. K2C translates nominal and ordinal variables, automatically fills in missing values and detects out of range data.

Benefits: Automated data preparation frees you to spend more time on model exploration and deployment.

KXEN Robust Regression (K2R) uses a proprietary regression algorithm to build predictive and descriptive models. These models can be used for scoring, regression, and classification. Unlike traditional regression algorithms, K2R can safely handle high numbers of variables (over 10,000). K2R provides indicators and graphs to ensure that you can easily assess the quality and robustness of your models.

Benefits: The data mining process is completely automated. The models provide drill-down into individual variable contributions.

KXEN Smart Segmenter (K2S) discovers natural groupings or clusters in a set of data. K2S is optimized to find clusters that are related to a specific business question. It describes the properties of each group and identifies how they differ from the general population. Like other KXEN modelling techniques it provides indicators for model quality and reliability.

Benefits: It automatically reveals the groups that are meaningful to the specific business questions you’re trying to answer.

KXEN Support Vector Machine (KSVM) is a binary classification component. It is particularly well suited for analyzing data sets with a small number of observations (rows) but with a high number of variables. This makes it ideal for problems in areas with very high dimensional feature spaces like life sciences.

Benefits: Problems that previously required customized programming can now be solved with this industrial strength software component.

KXEN Time Series (KTS) predicts meaningful patterns and trends in your data over time. Use your chronological data to forecast the results of the next periods of time. KTS identifies the trend as well as periodicity and seasonality to provide accurate and reliable forecasts.

Benefits: Adjust for patterns in your business and predict supply shortages before they occur.

KXEN Model Export (KMX) generates SQL, C, VB, SAS and other output code corresponding to the model built with the KXEN Analytic Framework. Models can easily be integrated into an application that supports these code types. KMX makes scoring independent from the modelling system and allows for very rapid deployment of models into production.

Benefits: KXEN models are rapidly integrated into databases, applications or business software without requiring the KXEN Analytic Framework. It also allows deployment of the models on platforms different from the one on which they were generated.

Copyright © 2002–2003 KXEN Inc. All rights reserved.