Data Technology and AI in Insurance – Unsupervised Learning & Recommendation Systems

Vocabulary flashcards covering the main concepts, techniques, and terms related to unsupervised machine learning, clustering, and recommendation systems discussed in the lecture.

Unsupervised Machine Learning

A family of algorithms that learn hidden patterns from unlabeled data, often by clustering similar items together.

Supervised Machine Learning

A learning approach where models are trained on labeled data that pairs inputs with desired outputs.

Cluster Analysis

A process in unsupervised learning that groups data points into clusters based on shared characteristics.

Clustering

The task of organizing data into clusters so that items in the same cluster are more similar to each other than to those in other clusters.

Hard (Exclusive) Clustering

A clustering method in which each data point can belong to only one cluster (e.g., k-means).

Soft Clustering

A clustering method where data points can belong to multiple clusters with varying degrees of membership.

k-means Clustering

A popular hard-clustering algorithm that partitions data into k clusters by minimizing distances to cluster centroids.

Proximity / Distance Measure

A quantitative metric used to determine similarity or dissimilarity between data points in multidimensional space.

Euclidean Distance

The straight-line distance between two points; best suited for dense or continuous data.

Manhattan Distance

The sum of absolute differences between coordinates; measures distance along axes like city blocks.

Cosine Similarity

A metric that measures the cosine of the angle between two vectors, indicating their directional similarity.

Granularity (in Clustering)

The level of detail determined by the number of clusters a user specifies for an algorithm.

Exploratory Data Analysis (EDA)

An application of unsupervised learning focused on uncovering patterns or structures in raw data.

Customer Segmentation

Dividing customers into groups with similar behaviors or characteristics, often using clustering techniques.

Cross-Selling Strategy

A business tactic that recommends complementary products or services, informed by clustering insights.

Anomaly / Fraud Detection

Identifying data points that deviate significantly from normal patterns, useful for spotting fraudulent insurance claims.

Recommendation System

Software that predicts and suggests items a user may like based on data analysis.

Collaborative Filtering

A recommender technique that predicts user preferences by analyzing behavior and similarities among many users.

Content-Based Recommendation

A recommender method that suggests items similar to those a user already likes, using item attributes for similarity.

Hybrid Recommendation System

A recommender approach that combines collaborative filtering and content-based methods for improved accuracy.

Trust in Recommender Systems

The user confidence that recommendations are helpful and unbiased—critical for long-term system success.

Labeling Cost

The financial and time expense of manually annotating large datasets, motivating the use of unsupervised methods.