With powerful AnswerTree, money can grow on trees

Which applicants are most likely to pay back loans on time?
What characteristics identify bad credit risks?
How can I reduce the number of bad loans?
What is the profile of customers who are most likely to go into default?

Decisions about extending credit to customers directly affect your organisation's bottom line. Today, simply grouping customers into "good" and "bad" categories is not enough. More in-depth analysis is required to effectively score an applicant's risk potential relative to the loan or line of credit.

Accurate and consistent decisions are crucial to your success. AnswerTree helps you create reliable customer profiles, identify customer groups based on behavior patterns and develop a model for making efficient and defendable credit decisions. AnswerTree gives you the power to make the most profitable credit decisions.

Make better decisions with more powerful models

AnswerTree is powerful decision-tree software that automatically identifies the unique characteristics of customers who are most likely to exhibit a certain action or behaviour, such as defaulting on a loan. Designed to be easy to use for both statisticians and non-statisticians alike, AnswerTree thoroughly analyses your data, discovers segments, builds models and predicts outcomes.

AnswerTree identifies risky customers or applicants using powerful segmentation techniques. By determining the natural groupings with similar characteristics, AnswerTree identifies the unique payment patterns, demographics, attitudes, and behaviours of your customers. Simply specify the target variable, such as status of credit account, and which variables to consider as predictors, such as monthly disposable income, credit limit, marital status, number of children, etc.

AnswerTree automatically sifts through your data finding statistically significant groups. With the knowledge of which groups are significant, you can then take action on what's most important. Your results appear in an intuitive tree diagram, colour-coded charts and tables, so it's easy to spot important relationships at a glance. Quickly identify segments to evaluate your customer profiles with precision.

It's easy to spot the best and worst segments at a glance. The results in this credit example reveal that credit limit is the best predictor of account status. AnswerTree found that customers with high credit limits exhibit the greatest percent (32%) of accounts that are 60 days late.

Gain more in-depth knowledge. Further results reveal more segmentation in the high credit customer group. Those customers with monthly disposable income of less than £968 had an even higher percent (51.65%) of the accounts that are 60 days late.

Discover significant segments. AnswerTree further identifies groups within the high credit customers with monthly disposable income of greater than £968. Of these customers, the majority (76%) of those with 2 or 3 children have Balanced accounts. Number of children is thus as an important predictor of account status.

Defend and be confident with your results

AnswerTree offers partitioning, so you can train your model on a subset of your data, then test the reliability of the model on the remaining portion. Or use cross-validation to test the performance of the model. The Misclassification table helps you quickly assess the accuracy of the model. You also get a value summary of each group on the Gains chart. By viewing the incremental costs and benefits of each node, you can carefully choose and exclude the appropriate groups, which means you save money and time.

Rely on the results of your model. The Misclassification table tells you the accuracy of your model, so you can make confident decisions. This credit scoring model accurately predicts 66 of 87 cases as 60 days late, for an accuracy rate of 76%. To get a more accurate model, AnswerTree offers powerful cost, profits and scoring that let you customise the model to incorporate what's important to you. For example, assign costs of £500 to 60 day accounts that are misclassified as Balanced, and AnswerTree uses cost complexity pruning to build a new model.

In addition, AnswerTree produces a model that you can easily apply to new cases. Save the decision rules to score new applicants. AnswerTree provides you with a logical framework for explaining to a customer why a loan was not approved. With AnswerTree, you get clear models, reliable results and a decision process that is explainable and defendable.

Credit risk analysis can be effective and efficient. AnswerTree provides you with the power you need in an easy-to-use, easy-to-learn tool that makes credit analysis more accurate, efficient and defendable. And, only AnswerTree offers four powerful algorithms for better segments and more reliable results. Identify which customers are likely to go into default and when to make collections efforts. Measure the probability of default more accurately. Save time and use a more objective decision process. With AnswerTree money can grow on trees.

See more on Answertree