This chapter focuses on individual univariate time series sampled uniformly through time. It describes the use of time-series features for tackling time-series forecasting. The chapter provides an overview of a vast literature of representations and analysis methods for time series. It explores global distances between time-series values, subsequences that provide more localized shape-based information, global features that capture higher-order structure, and interval features that capture discriminative properties in time-series subsequences. The chapter also describes feature-based representations of time series using the problem of defining a measure of similarity between pairs of time series, which is required for many applications of time-series analysis, including many problems in time-series data mining. Time series are a fundamental data type for understanding dynamics in real-world systems. The interdisciplinary reach of the time-series analysis literature reflects the diverse range of problem classes that involve time series. Global features refer to algorithms that quantify patterns in time series across the full time interval of measurement.