AI-Driven Market Analysis

Objectively evaluating data from a business and relating it to customized lists of merchant services. The technique involves using machine learning models to analyze business data and merchant service data, then deriving tailored recommendations of relevant services for each merchant. The models transform numerical representations of merchants and services into a common vector space to compare and correlate them. The models are trained on merchant profile data including statistics, categorical info, historical usage, and search history. This enables intelligent, near-real-time service recommendations based on aggregated merchant data.

Source

Demand sensing and forecasting system for supply chain using AI and big data processing to accurately predict product demand and optimize inventory levels. The system receives historical data, creates a hierarchical dataset, trains forecasting models, and uses a neural network to determine optimal weights for each level. It also considers factors like revenue, logistics, and political risks to provide recommended forecast frequencies. The system can dynamically process forecasts based on factors like engagement and pricing models, and categorize and classify products based on demand.

Source

Artificial intelligence (AI) powered real-time portfolio valuation system that accurately and efficiently values private company portfolios. It uses AI models to identify comparable public companies for each private company, retrieve financial data for those public companies, and then derive valuations for the private companies based on their identified comparables. This allows real-time portfolio valuation without subjective human estimation. The AI models use algorithms like boosting trees, k-nearest neighbors, and linear regression to cluster and compare companies.

Source

A method for building predictive analytics from text data extracted from earnings call transcripts using natural language processing (NLP) and machine learning. The method involves parsing the transcripts using NLP, creating intermediate metrics from the parsed text, combining the metrics into headline analytics, testing the headline analytics for standalone predictive power, selecting high-performing ones, then testing them again for additive predictive power when combined with existing market analytics. The selected headline analytics with additive predictive power are applied to new earnings call transcripts to predict financial performance.

Source

Dynamic forecast adjustment optimization using machine learning for collective optimization of an objective factor in a use-case like consumer goods. The method involves generating probabilistic forecasts, a relationship model, and an optimization model using ML systems. It recommends adjustments to collectively optimize the objective factor by dynamically generating recommendations based on the optimization model. This enables accurate forecasting and optimization across factors, objective, and constraints.

Source

Method for converting order book data into 2D format suitable for machine learning models. The method involves generating a tensor data structure with a time axis and price axis in tick units. The tensor has multiple 2D data items representing quantities for bid and ask prices at specific coordinates defined by time and price. A mid price calculated from the bid and ask is used as the tick price. This converts the tabular order book data into a format that can be directly input into 2D CNN models for stock price prediction.

Source

Improving the accuracy of market research data collection through autonomous evaluation and iteration. The method involves using machine learning models to identify potentially inaccurate collection information and then determining whether to obtain replacement data or generate simulated data to fill gaps. The decision is based on factors like event information and cost assessments. This allows cost-effective, iterative improvement of data collection accuracy.

Source

Using machine learning to efficiently analyze financial data and identify exceptions to financial algorithms. The method involves applying dynamic, interdependent algorithms to financial data using ML classifiers trained on labeled data. The classifiers classify outcomes as algorithm compliant, potentially non-compliant, or non-compliant. The ML allows efficient analysis of large datasets with complex algorithms by reducing computation and memory requirements. The classifier identifies latent anomalies in the financial data.

Source

Configuring AI decision nodes throughout a communication decision tree to dynamically define iteration data that corresponds to a trajectory through the tree. The decision nodes can support successive iteration of AI models to select the next node based on current profile data, learned data, and event detections. This iterative AI decision making can be availed through a canvas interface to visually define and connect the nodes, with each node representing an AI model or communication specification. The iterative AI selection allows customized and adaptive communication workflows based on real-time user context rather than static rules.

Source

A data analytics platform for forecasting commodity prices using machine learning and natural language processing on structured and unstructured data sources. The platform analyzes structured data like weather and futures prices, as well as unstructured sources like podcasts, to forecast commodity states. It uses techniques like neural networks, sentiment analysis, and time-series modeling to process the data and generate forecasts. The platform aims to provide accurate commodity price projections by leveraging multiple data sources and machine learning algorithms.

Source

Determining financial risk based on self-supervised natural language extraction from unstructured data sets like long form financial narratives. The method involves converting unstructured financial narratives into condensed financial risk narratives using self-supervised natural language processing. A tokenization library is determined for the unstructured data and used to generate the condensed narratives. Security scores indicative of financial risk are calculated for the condensed narratives. If a score exceeds a threshold, security actions are executed. This allows autonomous risk assessment from unstructured financial data without manual summarization.

Source

Dynamic demand forecasting system that provides accurate and interpretable forecasts of product sales at a granular level like individual product-location combinations. The system uses machine learning models trained on historical data to make forecasts. It dynamically selects the best model based on real-time conditions, or retrains existing models. The system also evaluates model performance using key metrics and performs inventory simulations.

Source

Online demand prediction using hierarchical temporal memory (HTM) for automated sales pipeline optimization. The method involves training an HTM network using historical demand data to learn temporal patterns representing sequences of states. This allows the HTM to forecast demand for offers by analyzing live data inputs in real-time without needing labeled training sets. The forecasted demand enables prioritization of sales of predicted popular offers.

Source

Generating feature embeddings for financial transaction data that can be used in machine learning models to improve prediction accuracy and interpretability. The embeddings are generated in an automatic and unsupervised manner using a heterogeneous graph representation of the transactions. The graph connects users and merchants based on transaction pairs with specific characteristics. The embeddings are generated hierarchically by jumping between transaction pairs based on frequency and hyperparameters. This allows extracting features with context and relationships between transactions. The embeddings can be labeled, indexed, and consumed by models without extra training.

Source

Training an AI model to predict the minimum winning price in real-time bidding without splitting auction types. The model is trained using auction data with conditional minimum winning price probability distributions. The overall minimum winning price probability distribution is generated first, then conditional distributions for each auction history. The AI is trained using auction attributes as input and the conditional distributions as output. This allows accurate price prediction in both auction types without dividing them.

Source

A system for facilitating electronic trading of financial instruments that uses machine learning to match traders and determine trade likelihood. The system allows users to initiate electronic trading sessions for specific financial instruments with specific terms. It matches the initiating user with dealer users who have a history of trading similar instruments. It also ranks potential invitee users based on likelihood of trading the instrument. This matching and ranking is done using machine learning models on historical trading data. The system then coordinates the trading session with selective sharing of offers.

Source

Intelligent valuation data reporting system that allows efficient and dynamic valuation analysis of startup companies using interactive graphical user interfaces. The system retrieves attributes of the companies and displays a customizable interface where the user can select thresholds for certain attributes. The system then dynamically populates a sub-interface with the companies that meet the selected thresholds, calculates valuations using AI models, and updates indicators for each company. This allows real-time, iterative, and customized valuation analysis and visualization of startup companies based on user-selected criteria.

Source

Automatically identifying counterpart entities from transaction strings in financial accounts when the counterpart entity is not explicitly listed. The method involves using machine learning models to translate the transaction strings into identified counterpart entities. The models are trained on manually mapped transactions and then retrained based on their predictions to improve accuracy. This allows automating the identification of counterpart entities from transaction strings that don't explicitly list them.

Source

Using AI and machine learning to predict and adjust demand for products and services across multiple channels like online, in-store, and delivery. An analytics platform with a machine learning network generates real-time and future demand metrics for each channel based on channel events. These metrics are then used to dynamically adjust prices and inventory allocations in each channel to optimize demand and availability.

Source

Method and system for high frequency stock trading that leverages machine learning models while avoiding issues like time gaps and latency. The method involves generating input data based on market data, feeding it to a machine learning model to predict stock prices at future times, and generating orders based on those predictions. To prevent time gaps, the input batch size is chosen based on latency to select a future time point before the batch arrives. This allows using the model without outdated predictions.

Source

A system for efficiently processing, organizing, and searching financial market data using vector representations. The system involves pre-processing the raw market data into vectors that capture the underlying structure and patterns. This involves filtering and transforming the data to extract meaningful features. The vectors can then be used for machine learning and analysis tasks. The search function allows finding similar historical market periods based on vector comparisons, enabling efficient exploration of related data.

Source

Selecting the optimal machine learning model for time series financial data with variable label distributions by training multiple models using oversampled training data with varying label distributions. During training, the label distribution is changed by oversampling specific labels. This generates a set of models with different label distribution biases. When making a prediction on new time series data, the similarity of its label distribution to the training distributions is measured and the most similar model is selected. This allows adaptive model selection based on the input data's label distribution.

Source

A method for providing dynamic, real-time data analysis and insights in large data sets without overwhelming processing resources. The method involves scanning the data to find the vital few contributors that account for most quality outcomes. By drilling down through the data to find the top results of profits, losses, conversations, etc., it captures a limited, manageable amount. This saved, live data enables dynamic analysis and visualization without affecting system performance.

Source

Predicting demand for products or services with improved accuracy by leveraging machine learning models that consider the distribution of actual demand values and the context around demand. The models estimate how the demand distribution shifts when certain variables deviate from their centers. This is done by learning from historical demand data paired with context variables like order volumes, weather, prices, etc. The models output predicted demand values and their confidence intervals at different levels of deviation from the center.

Source

Determining real-time personalized offers for users based on dynamic user attributes and data sets from similar users. The offers are dynamically determined in real-time based on constantly updated user attributes and user behavior. Machine learning algorithms analyze the user's dynamic attributes and similar user data to determine eligibility and likelihood of acceptance. This allows tailored offers that adapt to changing user circumstances.

Source

Real-time financial, inventory, and staffing optimization system for businesses using AI and machine learning. It leverages cloud-based predictive engines for finance, inventory, and staffing that analyze real-time data like location, time, user info, and third-party data to optimize decisions. It predicts patron attendance, staffing needs, inventory usage, cash flow, and generates recommendations for inventory adjustments, staff schedules, and loan risk profiles. The system integrates with user devices, databases, and gateways for vendors and financial institutions to optimize operations around business metrics.

Source

A natural language order fill module for trading systems that allows users to enter trades in free-form text without following a fixed ticket entry format. The module uses neural networks to interpret and extract trade details from unstructured input text, filling in the corresponding ticket fields in real-time. This allows users to enter trades as they would speak or type, with abbreviations, varying syntax, etc, instead of rigid ticket forms. The module also accepts copied/pasted text and learns ticket fields per product/asset class.

Source

Determining product roadmapping and strategy by gathering, organizing, structuring, and analyzing market, customer, competitor, and product data in real-time to make accurate and timely investment decisions and assess risks. The method involves obtaining data from users and sources, forming a taxonomy of classifications using models, selecting variables based on thresholds, evaluating outputs, and providing decisions and assessments.

Source

An automated trading platform that uses self-updating trading bots to find and deploy the optimal trading strategy over time. The platform generates a set of trading bots with parameters defining execution rules. It searches through the bots using historical data to find an initial ideal bot that tracks asset performance. This bot is deployed for a duration, evaluated, and then the search is repeated based on the initial bot's success. The platform deploys a new ideal bot for another duration. This iterative process allows the platform to dynamically adapt and improve its trading strategies over time using machine learning techniques.

Source

Mass execution of analytical models to predict optimal decisions in real-time for transaction negotiations. The system iteratively verifies data sources, generates transaction options, scores them using models, verifies compliance, and optimizes based on preferences. It determines a minimally viable transaction structure to constrain variation range. This parallel, asynchronous, non-blocking execution scales models across dimensions to quickly find optimal transactions.

Source

Automatically sensing, abstracting, perceiving, classifying, analyzing, and reporting regarding the actions of organizational entities in real-time or near-real-time using a networked, geographically distributed, entity sensing system. The system involves sensing entity actions using instruments, converting sensor signals to computer format, storing time/space/sensor data, pre-processing signals, abstracting actions, linking actions to regions, identifying entities, analyzing facts, generating reports, optimizing performance, and billing based on abstracted actions. The system uses machine learning to infer variables and relationships.

Source

Detecting abnormal trading activity using machine learning models that learn normal trading patterns for market participants. The models are trained on historical participant and external factor data using recurrent neural networks. They proactively monitor current trading patterns to identify deviations that pose risk, beyond normal thresholds given current conditions. The models use both participant and external data to identify abnormalities in real time.

Source

Multi-layer artificial intelligence framework for making dynamic predictions based on aggregated static and dynamic data over time periods. The framework involves using AI models that can learn and adapt predictions based on changing data over time, rather than just making static predictions based on data at a single point. It involves aggregating both static and dynamic features over multiple time periods to generate progressive predictions that reflect the evolution of data over time. This allows more accurate and dynamic predictions compared to static models.

Source

A method and device for predicting and controlling time series data using automatic learning. The method involves training multiple time series prediction models with different conditions, selecting optimal models that meet a criteria, and combining them into a final model. This allows leveraging the strengths of statistical and deep learning models. The final model is then used to predict target variables and control variables. It adjusts the control variables based on correlations with the predictions. The final model is evaluated, updated, or adjusted periodically to maintain performance.

Source

Simplifying complex market investments using AI-assisted trading recommendations and dynamic risk assessment. The platform uses AI to generate simplified trading recipes for complex investment opportunities based on factors like market sentiment, user preferences, and historical data. It ingests disparate sources of financial and sentiment data, analyzes it with machine learning, and packages investment opportunities with preconfigured trading recipes that match user risk profiles. This allows non-professional traders to execute complex strategies like options without deep knowledge or research.

Source

Automatically optimizing bid values in real-time bidding for resources like ad impressions by taking into account factors like uncertainty in user response probabilities and current bid budget and opportunity count. This allows more accurate and efficient bidding compared to static valuations. The method involves determining an adjusted value of the resource based on uncertainty and a risk tendency calculated from the bid budget and opportunities remaining. This adjusted value is used to determine the bid price.

Source

Data analytics system that can accurately forecast during crises when input data is volatile and inconsistent with historical data. The system uses a specialized crisis-recovery data analytics engine with a crisis-recovery-based machine learning model. The engine generates a consolidated single index of hierarchical and granular consumer behavior data during pre-crisis, crisis, and recovery phases. This crisis-recovery index allows accurate forecasting during crises when input data deviates from stable historical data.

Source

Using artificial intelligence and Data Shapley to develop decision-making models that can automatically adjust and reinvent themselves in real-time. The method involves decomposing a decision-making process into factors using data analytics and machine learning. It then uses Data Shapley to determine the most valuable new information to acquire. The process estimates utility functions for participants, analyzes trade-offs, and decompose the decision-making process to understand factor roles. This allows real-time decision-making with dynamic adaptation as conditions change.

Source

Demand prediction method that learns and adapts the prediction model by iteratively identifying and incorporating new external factors that influence demand. The method calculates the error between predicted and actual past demand, determines if it's abnormal, and if so acquires a new external factor. It then updates the prediction model with the new factor and repeats the process. This allows the model to continuously improve by discovering previously unknown demand drivers.

Source

Execution of stock orders using reinforcement learning to optimize trading strategies for individual stocks. The system collects historical stock data, trains a model using reinforcement learning to derive order execution strategies, and then executes orders in real time using the learned strategies. The reinforcement learning involves a model with multiple neural networks that determines action policies and estimates values for a reinforcement agent. The model is trained using stock data to derive strategies that maximize future rewards like trade execution cost. This allows the system to dynamically adapt and respond to market conditions for each stock when executing orders.

Source

Predicting interchange fees for credit card transactions in real-time to provide more accurate interchange fee estimates when processing transactions. The prediction involves using machine learning models trained on historical transactions to predict probabilistic distributions of possible interchange codes based on factors like card type, merchant type, submission timeliness, etc. These distributions are then used to calculate predicted interchange fees. By comparing predicted and actual fees, the models can be refined over time. This allows interchange fees to be estimated and held back from payments at the time of transaction processing instead of waiting for end-of-month statements.

Source

Method for more accurate demand forecasting using trend-informed neural networks to capture uncertainty in input trend forecasts and represent it in a latent space. The method involves training a model to capture uncertainty in input trend forecasts, representing it in a latent space using an auto-encoder, jointly optimizing the latent space and learning a time-series regressor from it, and then training the demand forecasting model using the optimized latent space. This allows more accurate demand forecasting that takes into account the uncertainty and variability of long-term trend forecasts.

Source

Determining accurate attribution weights for multiple touchpoints in a customer journey using a Markov chain model that optimizes order and sparse matrices to reduce data requirements and improve accuracy. The method involves: 1. Receiving customer journey data with touchpoints and channels. 2. Determining the optimal order for the Markov chain model based on the data. 3. Transforming the Markov chain transitions based on the optimal order. 4. Processing the customer data using a sparse matrix multi-level indexing model to generate sparse matrices. 5. Calculating removal effects and steady state values for the sparse matrices. 6. Determining attribution weights for the channels using the Markov chain model, removal effects, and steady state values. 7. Performing actions based on the attribution weights, like optimizing marketing spend or predicting future investments.

Source

Responsive stress testing system for analyzing market risk using machine learning and interactive visualization. The system has back-end ML models to compute market risk based on extracted features. A front-end interface allows loading models, specifying market shock ranges, and visually exploring how market variables impact portfolios. The ML calculations and visualization are distributed to the browser for scalability.

Source

Transaction data processing method for financial analysis using graph convolutional networks (GCNs) and transformers to extract spatial-temporal features from transaction graphs. The method involves obtaining transaction data for an account over multiple time windows, extracting spatial features using GCNs and temporal features using transformers, and generating a feature representation for the account based on the combined spatial-temporal information. This representation can be used for downstream analysis tasks like credit risk modeling, fraud detection, and money laundering detection.

Source

An intelligent and adaptive real time valuation engine using neural networks and regression analysis to provide accurate sub-millisecond valuations of products in highly fragmented, global, interlinked markets. The engine leverages LSTM neural networks to identify long term dependencies between data characteristics and transaction values. It receives input data from multiple sources, including continuous streaming channels, and assigns dynamic weighting to each predicted value based on the input data's influence. This adaptive weighting improves valuation accuracy in complex, interconnected markets with high volume and variety of data.

Source

Server and method for optimizing bidding prices in real-time bidding (RTB) advertising auctions. The server uses neural networks trained on historical data to predict winning probabilities for different bidding prices. It also takes user response probabilities into account. The server identifies the optimal bidding price based on the neural network outputs and user info. This allows advertisers to intervene in price determination rather than relying solely on statistical prediction models.

Source

Identifying trends in account behavior over time using embedding drift from graph neural networks. The technique involves generating embeddings for accounts based on transaction graphs at different time steps. It calculates the drift between the embeddings to measure change. By finding accounts with the highest drift, it can flag accounts with the most behavioral change. Repeating this over time provides a time series of normalized drift. Analyzing the shape of this trajectory reveals global trends across account groups. Clustering accounts with similar drift trajectories can group together accounts that follow similar transactional patterns.

Source

Dynamic pricing system for e-commerce retailers that uses dynamic optimization techniques to adapt prices in real-time based on demand and elasticity estimates. The system estimates demand using different models and estimates elasticity using a multi-armed bandit model. It then determines prices by optimizing revenue subject to constraints. This allows more accurate and dynamic pricing compared to static or passive learning approaches.

Source

Training a generative neural network to model market uncertainty and use it for portfolio optimization. The network learns the probability distribution of asset prices over time given past prices. This allows generating realistic future price trends. The network is trained on historical data. The probability distribution learnt by the network is then used to find diversified portfolios with optimal risk-return tradeoffs by simulating multiple market scenarios.

Source