Why to use it

The Feedforward Neural Network, also known as a deep feedforward network or a multilayer perceptron (MLP), returns a non-linear decision function. This works well when you have large dataset. 

If you only need text classification, try the Embedding attentive. If you are unsure of which model to use, the AutoLearn model can find the algorithm best suited to your data. 

How it works

The Feedforward Neural Network is a deep learning model. It is called feedforward because there are no feedback connections that feed output from the model back into itself. That is, connections between the network nodes do not form a cycle.

Historically, an early artificial neuron was called a perceptron. It took any number of inputs, added weights to express the importance of each input, and produced a single binary output of 0 or 1. Whether the output is a zero or a one is determined by whether the weighted sum of the inputs is greater or less than a threshold value. As you might guess, since another name for a feedforward neural network is multilayer perceptron, this deep learning model combines multiple layers of perceptrons or nodes.

So the Feedforward Neural Network by default uses three layers of nodes, each of which (except for the inputs) is a neuron that uses a nonlinear activation function:

  • an input layer,
  • a hidden layer,
  • and an output layer.

During training, the output values are compared with correct answers to create an error function, which is then fed back into the network to adjust the weights assigned to each input.

For more information on this concept, see Deep Learning: Feedforward Neural Network in Towards Data Science, and the Feedforward neural network article in Wikipedia. 

Several options for the activation function and optimization method are available.

Activation Functions

  • ReLU (default): A rectified linear unit function that sets all negative values to zero. This is the most commonly used activation function for deep learning.
  • Sigmoid: A logistic function with an S-shaped curve used to predict a probability as the output. This can cause it to get stuck during training, so it is not the default function.
  • Tanh: A hyperbolic tangent function, also with an S-shaped curve, used to classify into two classes. This is advantageous because negative inputs map strongly negative, and zero inputs map near zero in the graph.

Optimization Methods

Input fields

The model requires one Ground Truth field and at least one Unstructured Text or Training Feature field.

  • ground truth: a field that objectively measures how the user feels, such as a star rating
  • unstructured text: a field that collects free-form user feedback, that is, they type it rather than select from a list
  • training feature: a field that collects structured user data, that is, data that is numerical or that is selected from options 

Output fields 

The model returns the following fields for use in widget visualizations.

  • keywords: A textual array of the most important words within the data stream. In supervised models, keywords are the terms that the model finds and uses to predict each label. Keywords are only generated if you select an Unstructured text field. If you select only a Training Feature, that is, selectable values, dates, or numbers, no keywords are generated.
  • label: The predicted output based on what the model learned from your input data.
  • language.code: The two-letter language code of the detected language, e.g. en, fr, jp. See Language detection for more information.
  • language.name: The full name of the detected language, e.g. English, French, Japanese.
  • tokenized: A list of every word detected in the corpus. This is trivial for languages that use spaces between words, but for languages in which there are no spaces between words and multi-character words are possible, each requires a custom tokenizer.
  • unigrams: A textual array of single words within the data stream. Stratifyd calculates the total number of words and the number of unique values. Useful in a word cloud viewed with filters on average sentiment.
  • Data: A table containing all of the original data from the data stream, plus all of the analyzed data from the model.

To create a Feedforward Neural Network

You can create a model from within a dashboard, or you can add one to the Models page.

  • Create from dashboard
  • Create from Models page

Advanced settings

You can set the following properties in the Advanced section of the Create a new model or Deploy a new model dialog. 

  • Minimum Word Count: Set the minimum frequency of a given word to be considered for analysis.
    Default value: 2; Minimum value: 2
  • Resampling: If your data is skewed towards one class or another, you can resample your data and adjust the class distribution for better results.
    Default value: none; Valid values: Over sample, Under sample
  • K-fold Cross Validation: Set the number of equal-sized random subsamples to use for cross validation. The Ratio of training to validation set is ignored if you set this.
    Default value: 1; Minimum value: 1; Maximum value: 10
  • Ratio of training to validation set: The ratio that determines how much of the dataset is used as a training set and how much is used to validate the results. This is ignored if K-fold is set.
    Default value: 0.2 (80/20); Valid values: 90/10, 80/20, 70/30, 60/40
  • Minimum number of records for a field: Set the minimum number of records for a field to be considered for analysis. This prunes any classes with fewer records than the value you set here.
    Default value: 0; Minimum value: 0; Maximum value: 50
  • Activation Function: Select the nonlinear function to use in your model. See Activation Functions for more information.
    Default value: ReLU; Valid values: Sigmoid, Tanh, ReLU
  • Optimization Method: Select the learning optimization method to use in your model. See Optimization Methods for more information.
    Default value: ADAM; Valid values: SGD, ADAM, lbfgs
  • Run Language Detection: Run language detection on the source text.
    Default value: true
  • Default Language: Assume this language if language detection fails. This is used to select a language-specific stopword list to apply to clean text.
    Default value: English
  • Stopwords: Apply custom stopword lists to remove non-informative words from categorization.
    See Customize stopwords dictionaries for more information.
  • Chinese Dictionary: Customize Chinese tokens that our engine uses when creating key n-grams in your analysis.
    See Customize Chinese token dictionaries for more information.
  • Schedule Model Retrain: Set the period in days, weeks, months, or years at which to retrain the model.
    Default value: 0
  • Add Filter: Select a field on which to filter training data.

Did this answer your question?