How Ubcoin Market Determines the Price of Goods Using Artificial Intelligence

Ubcoin Market has developed a solution based on artificial intelligence to predict the optimal cost of goods. The system allows buyers to instantly receive objective information about the price of the goods in relation to the recommended price obtained on the basis of market analysis. Sellers are then able to confidently set an optimal price for their goods, ensuring demand.

Ubcoin Market is based on machine learning technologies in the field of text analysis, images and structured data. Ubcoin is conducting a public token sale, which has now raised more than 4 million dollars from private contributors and some undisclosed amounts from institutional ones — the Inventure fund and the Singapore-based Amereus fund. Ubcoin will be part of the Ubank mobile app that is pre-installed by Samsung in 10 countries and has over 16 mln installations worldwide. You can now buy discounted Ubcoin tokens on https://ubcoin.io/en

Most people have bought and sold goods through the Internet. When a person buys a product, one spends a lot of time comparing the price with other prices of similar goods. It is important for the seller to set such a price that will enable the goods to be quickly sold at an optimal price. Traditionally, the problem of establishing an acceptable price was solved by manual market research, which required a lot of time. In this day and age the buyer and the seller are not ready to waste time when it comes to a one-time purchase for a small amount. Another problem is that the average price is often not optimal, in some cases the seller will get more profit by lowering the price and selling more, or increasing the price and selling the product to another consumer segment of buyers.

A low price compared to the average is often a sign of counterfeiting, fraud, etc. Such ads and other similarly malicious ads require additional and effective moderation.

To solve these problems, the Ubcoin Market team began to develop a solution for forecasting the price of goods based on machine learning technologies. Machine learning technologies automate the process of price determination and provides better quality in forecasting compared with manual expert judgment. Ultimately, an expert cannot remember all the details of the goods and is subject to cognitive distortions, such as the binding effect. Whereas, our artificial intelligence is trained on all available examples of ads and analyzes and discovers nuanced data which is not obvious to manual human observations. Ubcoin Market uses information both about the setting of the price for the goods, and about the prices at which the goods are sold. Those variables are used to calculate an acceptable price for the seller with a minimal sale period.

Ubcoin Market Artificial Intelligence works as follows:

The AI collects information from ads regarding the specific price goods are listed at and the specified price goods are sold at.

Using the ad classification module, the subject matter of the ad is highlighted.

Then, the entities (name of the product, brand, product characteristics) are specifically identified.

Based on the received product characteristics, a forecast model is constructed that predicts the price of the goods by solving a regression equation with given inputs.

For the correct operation of the solution, the AI is required to collect a training sample consisting of several hundred ads for each product category. The classification of the ad topics is based on the determined classification of ad texts, and the determined classification of images, which is described in the article How Ubcoin Market uses machine learning to filter content. To classify images, a deep neural net is used based on the ResNet architecture. To classify the texts of rare declarations, the logistic regression algorithm uses the attribute vector received after applying the TF-IDF method over lemmatized, normalized words. To classify the texts of frequently occurring ads, a recurrent neural network is used.

After the selection of the named entities, a vector of characteristics is constructed, consisting of the characteristics of the goods with the value of the objective function, which corresponds to the price of the goods for which the goods are sold. To build the forecasting model, a boosting-based algorithm is used, supplemented by a linear model designed to predict the trend of price behavior.

For the client, it looks like this:

When creating an ad, the seller sees the recommended price, and the potential consumer’s level of interest in the product, depending on the price being set.

The buyer has the ability to see the price and interest level of other similar products.

The decision of Ubcoin Market to predict the price utilizing artificial intelligence helps the seller to set the optimal price for the goods and be confident in procuring demand. Furthermore, the buyer instantly receives information about the objectivity of the price of the goods. The use of artificial intelligence makes the marketplace for exchanging crypto currency for Ubcoin Market goods convenient and transparent for both buyers and sellers.