Auto-optimizing a neural net (and its architecture) on the CIFAR-100 dataset. Could be easily transferred to another dataset or another classification task.
This repository contains the implementation of a recurrent neural network (LSTM from keras library) with the purpose of forecasting target time series, given the targets historical records and covariates. The project uses a toy data set, while focusing on the data transformation tasks (pandas dataframes to 3D numpy arrays required by recurrent networks) and on the hyperparameters tuning tasks, taking advantage of keras_tuner package.
An exploratory data analysis is performed and a regression model is used to predict house values. The prediction performance is optimized after tuning the model hyper-parameters to minimize bias/variance errors.
