Energy Fraud Detection

Fraud detection services may be applied to different types of distribution, such as electricity or gas. With the aim of identifying irregularities or fraud on the installations, information provided by companies about users is analysed thoroughly: consumptions, meter locations, types of rates, etc.

For this analysis, predictive models and advanced techniques of variables selection and generation are used, and the most relevant techniques regarding the problem raised are identified using correlations, statistic indicators, relations among clients, QPFS or mRMR, among others. Then classification techniques are applied using historical data both from clients’ behaviour and structural data.

On the basis of this data, the system updates information about clients periodically —according to the needs of the distribution company—, and provides a ranking including the score assigned to every client. Besides, the information contained on the data, and possible interpretations, may be represented graphically to understand it better. Thus after the analysis is performed, a group of clients that might be committing fraud against the company is selected and shown.

A company may use this information to organise visual checks of the installations, the result of which will also permit training predictive models again.

Detecting fraud among users early allows energy distribution companies to reduce losses associated with fraudulent use of resources. Besides, results from analyses permit optimising the number of checks. This way, investment on the service may be redeemed on the short term, avoiding costs incurred by unnecessary checks.

IIC team of experts uses own technology that can be adapted, enhanced and renewed according to specific needs for every problem. Thus, the system can be adapted to several kinds of energy distribution and different geographical areas inside and outside Spain, always guaranteeing the best model of fraud detection depending on particular needs of every client.