GraphConnect 2018

In our work, we combine the technology of IBM’s AI, Watson Assistant, with the use of graph databases supported by Neo4j. We solved the problem of making a recommendation to a customer in the shortest time possible, based on certain questions that the bot asks the user. The customer will feel comfortable, as they will be able to interact with the Chat-bot in natural language, and receive a response based on their interests or requests in a short time. This type of solution can be adapted to any type of recommendation or search that the customer desires. The graph database will contain all the information necessary to make the recommendation, may that be a movie database (title of the movie, names of the actors, directors, genre, year of release, etc.), the catalogue of certain products in a store (say, all refrigerators with their capacities, color, size, electric efficiency, etc.), or a database of people with certain information about them (occupation, age, place of birth, favorite meal, etc.). The graph will be provided with weights on certain edges, which we calculate using a metric of our choice. These weights will be used in order to traverse the graph in the most efficient way, and to, in that way, get to a certain node (or set of nodes) that we will present the user as our recommendation.