Replication Data for: Predictive Modeling of Concrete Stress-Strain Relationship Using P-LSTM

This repository contains the implementation of the Pruning Long Short-Term Memory (P-LSTM) model for predicting the stress-strain relationship of normal weight aggregate concrete (NWAC) and lightweight aggregate concrete (LWAC) at various temperatures. The code and data included here are designed to support the findings of the research article "Pruning Long Short-Term Memory (P-LSTM): A Model for Predicting the Stress-Strain Relationship of Normal and Lightweight Aggregate Concrete at Finite Temperature." The repository provides a comprehensive suite of tools for preprocessing data, training the P-LSTM model, and evaluating its performance against experimental results. The model leverages advanced deep learning techniques to improve prediction accuracy and computational efficiency, making it a valuable tool for civil engineering applications, particularly in the design and optimization of concrete structures. The P-LSTM model has shown significant improvements over traditional LSTM networks in terms of accuracy, stability, and computational cost. This makes it especially useful for applications where data size and computational resources are limited. The provided datasets include experimental measurements of concrete mixtures subjected to varying temperatures, and the scripts offer a detailed workflow for replicating the study's results. Files Description: data.xlsx: Contains the input data matrix where each row represents a data point and each column represents a dimension of the data. out_PY_T1_1.xlsx to out_PY_T4_30.xlsx: These files contain the output data from the experiments categorized by different temperature exposures and mixture designs. Each file represents the results from a specific set of conditions used in the study.