Install
Predictive terms
Forecasting, machine learning and scoring: what will happen.
ARIMA (autoregressive integrated moving average) A classical time-series forecasting model that combines a series' own past values, differencing, and past forecast errors. AutoML (automated machine learning) Tooling that automates parts of building a machine learning model, such as feature selection, algorithm choice and hyperparameter tuning. Classification A supervised learning task that assigns each input to one of a fixed set of discrete categories or labels. Clustering An unsupervised technique that groups data points so items within a group are more similar to each other than to items outside it. Computer vision A field of AI that trains models to interpret and extract information from images and video, such as detecting objects or faces. Concept drift A change over time in the underlying relationship between a model's inputs and the outcome it predicts. Confusion matrix A table comparing a classifier's predicted labels against actual labels, breaking results into true and false positives and negatives. Cross-validation A method for estimating how well a model generalizes by repeatedly training and testing it on different splits of the same data. Decision tree A supervised model that predicts an outcome by following a sequence of if-then splits on feature values, shaped like a tree. Deep learning Machine learning using neural networks with many layers, able to learn complex patterns directly from raw data such as images or text. Demand forecasting Estimating future customer demand for a product so inventory, production and staffing can be planned ahead. Dimensionality reduction Techniques that compress a dataset with many variables into fewer dimensions while preserving as much meaningful information as possible. Displacement risk A modeled estimate of how likely a neighborhood or household is to experience housing-driven displacement. Elo rating A rating system that updates a competitor's skill estimate after each result, weighted by the strength of the opponent faced. Ensemble forecasting Running many slightly varied forecast simulations to express weather or climate uncertainty as a range instead of one number. Exponential smoothing A forecasting technique that weights recent observations more heavily than older ones, with the weight decaying exponentially with age. F1 score The harmonic mean of precision and recall, giving a single score that balances both when neither error type is clearly more costly. Feature engineering Creating, transforming or selecting the input variables a model is trained on, to make the underlying pattern easier for it to learn. Gradient boosting An ensemble method that builds trees one at a time, each one correcting the errors made by the trees before it. Hyperparameter tuning The process of finding the settings that control how a model learns, such as learning rate or tree depth, rather than what it learns. K-means clustering A clustering algorithm that partitions data into k groups by repeatedly assigning points to the nearest of k cluster centers. Likely voter model A statistical filter pollsters apply to a survey sample to estimate results among the people who will actually vote. Load forecasting Predicting future electricity demand so utilities can plan generation, purchasing, and grid operations. Logistic regression A statistical model that estimates the probability of a binary outcome as a function of input variables. Machine learning Building models that learn patterns from data to make predictions or decisions, rather than following hand-coded rules. Microtargeting Using granular voter or consumer data to reach narrowly defined audience segments with tailored messages. Monte Carlo simulation A technique that models uncertainty by running a scenario many times with randomly sampled inputs to see the range of outcomes. Neural network A model made of layers of connected nodes that learns to map inputs to outputs by adjusting connection weights during training. Overfitting When a model learns the noise and specific quirks of its training data so closely that it performs poorly on new data. Persuasion modeling Predicting which voters are movable on a candidate or issue, as distinct from how likely they are to vote. Population health management (PHM) Using aggregated patient data to improve health outcomes and control cost across a defined group of people, not just individuals. Precision and recall Two complementary classifier metrics; precision is how many predicted positives were correct, recall is how many actual positives were found. Predictive analytics Analytics that uses historical data and statistical models to estimate what is likely to happen next. Predictive maintenance Using sensor data and models to predict when equipment is likely to fail so it can be serviced before it breaks down. Predictive policing Using historical crime data and statistical models to forecast where or by whom future crime is likely, to guide patrols. Propensity model A predictive model that scores how likely an individual is to take a specific action, such as buying, churning, or clicking. Random forest An ensemble model that trains many decision trees on random subsets of data and features, then averages their predictions. Regression analysis A statistical method for estimating the relationship between one outcome variable and one or more predictor variables. ROC AUC (area under the ROC curve) A single number summarizing how well a classifier separates positive from negative cases across every possible decision threshold. Sales forecasting Predicting future sales volume or revenue from pipeline data, historical trends or statistical models. SHAP values (SHapley Additive exPlanations) A method for explaining an individual prediction by fairly attributing it to the contribution of each input feature. Small area estimation (SAE) Statistical techniques for producing reliable estimates for areas or subgroups too small to sample directly. Species distribution model (SDM) A statistical model that predicts where a species can occur by relating known sightings to environmental variables. Supervised learning A machine learning approach where a model learns from examples that already include the correct answer, to predict that answer on new data. Time series forecasting Predicting future values of a metric using patterns found in its own historical sequence over time. Topic modeling An unsupervised machine learning technique that discovers the latent themes running through a collection of documents. Train-test split Dividing a dataset into separate portions for training a model and for evaluating it, so performance is measured on unseen data. Training data The dataset with known outcomes that a machine learning model learns patterns from before being used to predict new cases. Turnout modeling Predicting the probability that an individual registered voter will cast a ballot in a given election. Unsupervised learning A machine learning approach that finds structure or patterns in data that has no labeled correct answer to learn from. Uplift modeling Predictive modeling that estimates how much a treatment changes an individual's outcome, to target only the people it will actually influence. Win probability model A statistical model estimating a team's real-time chance of winning a game given the current score, time remaining and situation.