Features

Easy to use and efficient means for risk analysis optimized for personal loans include detection of:


  1. •Fraud signs

  2. •Client difficulties signs

  3. •Technical risks for debt recovery.


Ease of Use:


  1. •Analysis scenario and know-how is included into the product and does not require special knowledge and math skills from the user

  2. •Generates concise report and output data spreadsheet

  3. •Risk factors detection rules are based on easy to understand “IF ... THAN ...” yet powerful fuzzy logic

  4. •Uses single data input file: flat spreadsheet in text (CSV) format

  5. •Requires minimal data set for analysis and learning, does not imply extra integration efforts

  6. •Can work in batch mode e.g. for scheduled loans portfolio monitoring.


Efficiency:


  1. •Fraud detection and default risk factors knowledge base uses fuzzy logic with adaptation to input data

  2. •Behavior statistic is used for analysis

  3. •Analysis is performed from Collector point of view (in contrast with traditional risk analysis approach based mainly on demography statistics)

  4. •Concise reports are easy to understand and provide in-depth risks monitoring from different angle.


The Result:


  1. •Easy to understand concise report: accurate and unbiased decision maker’s aid

  2. •Debtors spreadsheet for integration into IT infrastructure contains detailed risk factors definitions and quantitative measures

  3. •Daily data processing time economy (up to 100 man-hours economy for 1000 debtors portfolio compared to manual analysis).


Find out more...


Unlike traditional data analysis software Collect Advantage does not require math skills from the users. SAS, SPSS and other popular statistical analysis packages include tons of data mining tools for professional researchers. Our approach is different: provide concise, easy to learn vertical market solution for everyday practical use.

Fraud detection



Fuzzy search algorithms are used everywhere in Collect Advantage to find hidden relationships between debtors. These relationships are graphically displayed with point and click functionality.




Fuzzy rules are widely used for risk evaluations. They naturally fit into imperfect data analysis system and can efficiently absorb data obtained from fuzzy text search algorithms.

Easy to use tool for confident prognosis of debt recovery and bad loans evaluation:


  1. •Cash flow prognosis of debt recovery

  2. •Accounts segmentation

  3. •Debt portfolio evaluation


Ease of Use:


  1. •“One-click” operation does not require special knowledge or math skills from the user

  2. •Data processing procedures and know-how are included

  3. •Generates concise report and output data spreadsheet

  4. •Includes means for prognosis parameters adjustment and neural network training

  5. •Uses single data input file: flat spreadsheet in text (CSV) format

  6. •Requires minimal data set for analysis and learning, does not imply extra integration efforts.


Efficiency:


  1. •Prediction method is Neural Network

  2. •Neural Network inputs are fuzzy set values

  3. •Innovative mathematic model for time-based cash-flow prediction

  4. •Classification error below 15%, cash flow prediction error below 10% for 1000-debts packs.


The Result:


  1. •Easy to understand concise report: accurate and unbiased decision maker’s aid

  2. •Debtors spreadsheet for integration into IT infrastructure contains detailed risk factors definitions and quantitative measures

  3. •Time-based cash flow prediction for every account and for total loans portfolio

  4. •Bad debts segmentation for outsourcing or selling to collector agencies

  5. •Loans portfolio pro-active monitoring

  6. •Lowering bank costs of personal loans management

  7. •Get rid of bad loans early

  8. •Concentrate on good loans recovery.


Find out more...

Collection Scoring



Neural nets are best working when input data is not ‘real world’ measurements but credibility measures of a value belonging to some class. This makes fuzzy sets the ideal data pre-processing technique.


Intelligent system for debt recovery management optimization:


  1. •Debt recovery plan (strategy) recommendation for each particular account

  2. •Collection costs vs. recovery prognosis based optimizing

  3. •Efficient resources allocation for collection procedures


What it does:


  1. •System recommends optimal recovery strategy for each account

  2. •Recovery strategy is ‘communication mix’ e.g.:

  3. •intensive phone calls (daily)

  4. •SMS with weekly phone calls

  5. •hard collection

  6. •legal actions

  7. •Costs apply:

  8. •communication mix costs

  9. •percentage costs calculated per account (legal fees, discounts, outsourcing costs)

  10. •taxes and other hidden costs applicable (e.g. cost of funds securing)

  11. •Business limitations apply:

  12. •resources

  13. •priorities

  14. •coverage

  15. •etc.


Efficiency:


  1. •Well established math method of linear optimization ensures sound results

  2. •Flexible linear problem setup by means of GNU MathProg language.


The Result:


  1. •Automatic recovery strategy selection for every account based on cash flow prognosis and collections cost

  2. •Collection resources optimal allocation

  3. •Notable effect when implementing at medium-to-large collection departments e.g.:

  4. •50+ employees call centers

  5. •nationwide networks of hard collection and legal collection departments

  6. •debt portfolio of 1,000+ clients


Find out more...

Automatic Strategy Selection