FNNS: An Effective Feedforward Neural Network Scheme with Random Weights for Processing Large-Scale Datasets

Abstract

The size of datasets is growing exponentially as information technology advances, and it is becoming more and more crucial to provide efficient learning algorithms for neural networks to handle massive amounts of data. Due to their potential for handling huge datasets, feed-forward neural networks with random weights (FNNRWs) have drawn a lot of attention. In this paper, we introduced an efficient feed-forward neural network scheme (FNNS) for processing massive datasets with random weights. The FNNS divides large-scale data into subsets of the same size, and each subset derives the corresponding submodel. According to the activation function, the optimal range of input weights and biases is calculated. The input weight and biases are randomly generated in this range, and the iterative scheme is used to evaluate the output weight. The MNIST dataset was used as the basis for experiments. The experimental results demonstrate that the algorithm has a promising future in processing massive datasets.

Fundamentals, Algorithms, and Technologies of Occupancy Detection for Smart Buildings Using IoT Sensors

“…RNNs are designed specifically to handle sequential data, where the timing and order of the information are crucial. Recurrent connections, not present in FNNs or CNNs, enable information to pass from input to output and through loops that feed data from earlier time steps into the current step [65,83]. This qualifies them for jobs like time series analysis, speech recognition, and natural language processing.…”

“…This qualifies them for jobs like time series analysis, speech recognition, and natural language processing. RNNs can recognize temporal connections and context in the data since they have a memory of past inputs [83]. However, due to problems with disappearing gradients in lengthy sequences, advanced variations, such as Long Short-Term Memory (LSTM) [84] and Gated Recurrent Unit (GRU) networks, have been developed.…”

“…• In occupancy detection, the input layer receives sequential data representing features at different time steps [65,83]. • Input features: [X = x 1 , x 2 , .…”

“…• The recurrent layer in RNNs is designed to capture temporal dependencies in sequential data. • It keeps hidden states constant over time, enabling the network to recall data from earlier time steps and apply it to predictions at the present step [65,83].…”

“…• The output layer of an RNN is in charge of making predictions, much like the output layer in a Feedforward Neural Network (FNN) [65,83].…”