Jet Infosystems developed for network of drugstores "A century live" a system based on algorithms of machine learning

On October 15, 2018 the Jet Infosystems company announced development for network of drugstores "the CENTURY LIVE" a system based on algorithms Machine Learning (ML) for commodity recommendations at the checkout. The solution allows pharmacy chain to make more address recommendations to buyers and to raise an average bill.

The developed service obtains data on purchases of the client, analyzes them using model of machine learning and sends to the employee at the checkout the list of the recommended goods. Enter it TOP-3 medicines which the client with high probability will add to the purchases when providing the qualified recommendation. The service offers specific goods items to within the article, selecting them from 27 thousand names of medicines and parapharmaceutical products. The automated recommendations are designed to help with the organization of additional sales without creation of essential load of the staff of pharmacy chain. All work on determination of the goods, most interesting to each certain buyer, is undertaken by the software solution.

The automated ML tool helps to reveal, on the one hand, the hidden customer needs, and with another — to provide them useful and rather exact recommendations about the individual list of the purchased medicines. Task of this solution: increase in the average check at the expense of address, necessary to the buyer, offers.

For drawing up individual forecasts specialists Jet Infosystems used the whole complex of methods of machine learning. The mathematical model of service was trained on data from check storage of the company for the long period. Here the structure of the check, the list of the purchased goods and the price on each position entered. To the solution opened access and to information on the directory of medicines with breakdown on separate categories of medicines. On the basis of the analysis of an extensive pool of earlier made purchases ML algorithms are capable to recommend additional offers on each check passing through a system with rather high accuracy.