Gene Expression Programming in Data Mining Applications

Gene expression programming (GEP) is an evolutionary model that combines the advantages of genetic algorithms and genetic programming to evolve computer programs or mathematical expressions for data-driven tasks. In data mining, GEP excels at function discovery, classification, regression and correlation mining by evolving populations of candidate solutions encoded as linear chromosomes that translate into expression trees. This dual representation supports efficient search and flexible model structures, enabling the automatic synthesis of predictive models without manual feature engineering. Recent advances have addressed key challenges such as convergence speed, computational scalability and interpretability. Enhanced variants filter irrelevant features, exploit schema theory to guide evolution, or leverage parallel and distributed computing frameworks to handle large-scale, high-dimensional datasets. Applications span industrial system diagnostics, bioinformatics, finance and environmental monitoring, where GEP uncovers nonlinear relationships and yields human-interpretable expressions. The global significance of these developments lies in providing adaptable, transparent and scalable tools for complex pattern discovery across diverse domains.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

Event-tracker enhanced GEP frameworks have been introduced to pre-filter data components in complex industrial systems, reducing search space and accelerating convergence while retaining model accuracy. By integrating event-tracking mechanisms into chromosome evaluation, these approaches significantly shorten runtime for correlation mining problems. Schema theory-based engineering of GEP has provided a principled basis for parallel implementation: by aligning chromosome generation structure with data segmentation strategies, large-scale analytics on multi-core or cluster environments achieve substantial speed-ups without sacrificing solution quality. In parallel classification tasks, GEP algorithms have been adapted to distributed computing models, such as MapReduce, to enable efficient function mining on large datasets; experiments demonstrate linear scalability and robust performance on high-dimensional real-world benchmarks.

Gene Expression Programming in Data Mining Applications publication trend

The graph below shows the total number of articles in gene expression programming in data mining applications across all publications each year (not limited to Nature Index journals).

Technical terms

Gene expression programming (GEP): An evolutionary algorithm that encodes solutions as linear chromosomes, which are expressed as tree-like programs for optimisation or modelling tasks.

Chromosome: A fixed-length linear representation of a candidate solution in GEP, comprising genes that map to subtrees in the expression tree.

Schema theory: A framework describing how patterns (schemata) are propagated through generations in evolutionary algorithms, guiding the design of genetic operators.

Function mining: The process of discovering mathematical expressions or programs that capture relationships in data.

Parallelisation: The distribution of computational tasks across multiple processors or nodes to improve runtime performance.

References

1. EGEP: An Event Tracker Enhanced Gene Expression Programming for Data Driven System Engineering Problems. IEEE Transactions on Emerging Topics in Computational Intelligence (2019).
2. Schema Theory-Based Data Engineering in Gene Expression Programming for Big Data Analytics. IEEE Transactions on Evolutionary Computation (2017).
3. Parallelizing Gene Expression Programming Algorithm in Enabling Large‐Scale Classification. Scientific Programming (2017).
4. Distributed Function Mining for Gene Expression Programming Based on Fast Reduction. PLOS ONE (2016).