Improving the performance of a neural network

Techniques to improve the performance of a neural network

In this article, we will explore how to identify if we have an under-fitting or over-fitting neural network and then apply appropriate techniques to improve the performance of the neural network

Features of neural network

Objective of a model built using neural network is to perform well on training data by generalizing. This helps to make better predictions on unseen data in test set or validation set.

Neural networks

• Mimic human brain
• Trained using fast enough computers with good amount of data, structured or unstructured. Data can be tabular data, images, text files, audio and video files
• Perform automatic feature extraction
• Have scalable performance. Perform well on training data as well as unseen data that is not part of the training data
• Low error on training data set as well as low error on test or validation data set
• Find the optimal weights for the neural network to get accurate results

General techniques to improve the performance of a neural network