When thinking about machine learning, how often do you consider time? Time series analysis is an important area of machine learning because many predictive problems that ML is used to solve involve a critical time component.
Often, solving these predictive problems is sometimes neglected because the time component makes these problems more difficult to handle and process. Time series analysis is the method by which such problems can be effectively solved.
Below, we’ll discuss what time series analysis is and some practical, real-life applications.
What is time series analysis?
A time series is a series of data points indexed in time order. The daily values of the Dow Jones Industrial Average and a car’s engine temperature measurements taken from a sensor are both examples of a time series. Time series analysis includes methods for analyzing these series of data to extract meaningful statistics, predict future values, or detect anomalies.
Time series analysis comes in several forms, including:
- Descriptive analysis: determining trends and patterns, such as “cyclic patterns, turning points, and outliers”.
- Spectral analysis: separating and determining periodic or cyclical components in a time series.
- Forecasting: predictions on the future based on historical trends.
- Intervention analysis: determining if a specific event has lead to a positive or negative change in the time series.
- Explanative analysis: determining the relationship between two or more time series analyses.
A variety of different techniques are used to carry out time series forecasting, including:
- Multilayer perceptron.
- Bayesian neural networks.
- Radial basis functions.
- Generalized regression neural networks (also called kernel regression).
- K-nearest neighbor regression.
- CART regression trees.
- Gaussian processes.
What are the benefits?
In a nutshell, time series analysis is most often instrumentalized for its predictive function, since it is capable of authoritatively predicting future trends and changes. This predictive power can be leveraged by a variety of practical applications. Stock price prediction, equipment malfunction detection, and cyber attack monitoring are a few key examples of how time series analysis can be applied to real-life scenarios. Other applications include:
- Meteorology: weather variables, like temperature, pressure, wind.
- Economy and finance: economic factors (GNP), financial indexes, exchange rate, spread.
- Marketing: activity of business, sales.
- Industry: electric load, power consumption, voltage, sensors
- Biomedicine: physiological signals (EEG), heart-rate, patient temperature.
- Web: clicks, logs.
- Genomics: time series of gene expression during cell cycle.
The value of a time series framework?
A typical time series workflow usually requires multiple iterations, trying different models and algorithms, tweaking their parameters and applying them in many different combinations. As a result, it is a manual, time-consuming, and complex process.
At Digamma.ai, our team has made this development process better, faster and more efficient by developing our own proprietary time series analysis framework.
The framework consists of four parts including data preprocessing, exploratory analysis, anomaly detection and forecasting. Each of these four methods includes a variety of models and algorithms each customizable with options and parameters. In short, our framework completely automates the traditional time series workflow.
Our approach allows for the use of a variety of effective methods quickly, iteratively and easily to refine a time series model instead of having to rely on onerous trial and error experiments.
Importantly, our framework is scalable to very large datasets and can be deployed to the cloud, including Amazon AWS and other providers.