Time Oriented Language, hereafter TOL is a programming language designed for analysis of time series and stochastic processes based on the algebraic representation of time and the time series which allows:
- Give structure to the data from operational systems, giving them a temporary support and classifier, which makes them useful information to understand their behavior.
- Analyze the dynamic information, generate statistical models, to identify factors that influence the timing behavior and extract knowledge.
- Facilitate decision making knowledge inferred from statistical models, forecasts of behavior, and decision functions.
Time is a continuous magnitude represented as the real line, but when dealing with time series of social nature (sales, phone calls, bank transactions, event attendance, traffic, etc.), the analyst notes on behaviors involving different temporal characteristics such as cycles within a period (which may be the day, week, month, year ...); or proximity to holidays, bridges and any other identifiable events when they occur.
Traditionally, these phenomena have been represented by extension, that is, as drawn up a list of events more or less handmade, which is not a very serious problem for annual or monthly series. In recent years, the information explosion has brought a series of daily data, hours or even higher frequencies, for which necessitates an algebraic treatment of this social time. Based on a temporary basis sets primary and a collection of functions, enable analytically build complex expressions to represent the most intricate social behaviors.
On the other hand, the language must provide mechanisms for statistical analysis of data from reading them in files or databases to produce predictive models, through the methods of visual representation of information that may be necessary.
Within this context TOL born to try to answer all these needs and others that have arisen during the implementation process and maturation of language.
TOL was not possible without the contribution of next open source projects (alphabetical order):
- ALGLIB: Efficient multilingual scientific software library
- ANN: ANN: A Library for Approximate Nearest Neighbor Searching
- ATLAS: Automatically Tuned Linear Algebra Software
- BLAS : Basic Linear Algebra Subprograms
- BZip2: High-quality data compressor
- CHOLMOD: Sparse Cholesky factorization and update/downdate
- CLUSTERLIB: The C clustering library for cDNA microarray data
- DCDFLIB: Evaluating cumulative probability density functions
- Google Sparse Hash: An extremely memory-efficient hash_map implementation
- GSL: GNU Scientific Library
- IPOPT: Interior Point OPTimizer a software package for large-scale nonlinear local optimization
- KMLOCAL: Efficient Algorithms for K-Means Clustering
- LAPACK: Linear Algebra PACKage
- NLopt: Steven G. Johnson, The NLopt nonlinear-optimization package
- optimal_bw: Fast O(N+M) Univariate Kernel Density Estimation with Optimal Bandwith
- R: The R Project for Statistical Computing
- ZipArchive: Adds ZIP compression functionality to your software