Automated order execution based on user preference settings utilizing a neural network prediction model

A computer-implemented method comprising:

• receiving, by an online application associated with a user device, user preference information, and transmitting the user preference information to a server;
• determining, by the server, one or more user settings based on the received user preference information, wherein the one or more user settings comprises a price limit associated with the user preference information;
• training, by the server, a neural network prediction model using back propagation based on a plurality of alternate data inputs of different types associated with at least one company;
• applying, by the server, the neural network prediction model to new data from a plurality of data sources to generate a forecasted value of a critical indicator of the at least one company;
• determining, by the server, a data condition based on the forecasted value of the critical indicator of the at least one company;
• providing, by the server, an electronic notification indicative of the data condition to the user device based at least on the price limit of the one or more user settings, wherein the electronic notification includes a recommendation associated with the at least one company based on the data condition, and a query to perform an automated user request in accordance with the recommendation;
• and
• executing, by the server, the automated user request in response to an automated determination to perform the automated user request based on the one or more user settings, wherein executing the automated user request comprises automatically submitting an order, in accordance with the recommendation, to a brokerage service without further input from the user,
• wherein training the neural network prediction using model back propagation comprises:
• receiving, by the server, the plurality of alternate data inputs of different types from a plurality of alternate data providers, each of the plurality of alternate data providers having different types of data;
• standardizing, by the server, the plurality of alternate data inputs of different types received in real-time from the plurality of alternate data providers by performing filtering, de-duplication, normalization, and classification;
• receiving, by a modelling system for the neural network prediction model executed by the server, the standardized plurality of alternate data inputs of different types and published information of the at least one company, the modelling system employing neural networks to generate, as an output, the forecasted value of the critical indicator of the at least one company;
• receiving the output as the forecasted value of the critical indicator of the at least one company in real-time, comprising one or more revenue predictions;
• and
• performing, by the server, model validation using historical data associated with the at least one company using back propagation to periodically update the neural network prediction model;
• for the neural network prediction model being determined as not successfully validated, iteratively performing, by the server, retraining and updating of the neural network prediction model until the neural network prediction model is successfully validated.