Introduction

This module reference is intended to provide an inventory of the technical and functional background of the modules available within the SmartPredict toolbox.

Each module is a component that fulfills a well-defined task in the model. Their user-friendly layout will ensure that you will have a fluid, out-of-the-box user experience.

Simply drag and drop them into the workspace in order to assemble your Machine Learning or Deep Learning model flowcharts.

Where are they located?

Core modules can be accessed from the project workspace. Within the right sidebar, there displays an inventory of these available modules . To extend your project's scope, it is moreover possible to create custom modules through coding with the help of the Notebooks. For now, we are only going to focus on the description of these core modules.

Designing a flowchart , be it for a build or for a deployment always follows an easy principle: simply drag and drop modules into the workspace in order to obtain a viewpoint of the data flow.

Various, versatile or specific, the core modules cover a large number of AI use cases, as simple as classification projects to more complex Natural Language Processing, passing through Computer Vision.

Module organization

SmartPredict' s core modules are gathered into 15 functional categories . The modules that fulfill similar kinds of tasks are located under the same sub-tab. They may share common libraries or algorithms .

Modules sometimes differ in terms of purpose: for instance , one is a classifier and one is a regressor. Often, modules can be used interchangeably in order to find the right fit for the models.

Module reference

Here is the exhaustive list (up to this version but not limited to this) of SmartPredict' s palette of modules according to their function.

  1. Basic Operations

    1. Item Saver

  2. Web Service

    1. Web Service IN

    2. Web Service OUT

  3. Data Retrieval

    1. Data Fetcher

    2. Data Frame loader/converter

    3. Image Data Loader

  4. Data Preprocessing

    1. Array Reshaper

    2. Generic Data Preprocessor

    3. Missing data handler

    4. Normalizer

    5. One Hot Encoder

    6. Ordinal Encoder

  5. Data Selection

    1. Features selector

    2. Generic data splitter

    3. Labeled data splitter

  6. Training and Prediction

    1. Predictor DL models

    2. Predictor ML models (Probabilistic models)

    3. Trainer ML models

    4. Trainer/Evaluator DL models

  7. Evaluation and Fine Tuning

    1. Cross Validator for ML

    2. Evaluator for ML models

  8. Machine Learning Algorithms

    1. KNeighbors Classifiers

    2. KNeighbors Regressor

    3. Linear Regressor

    4. Logistic Regressor

    5. MLP Classifier

    6. Random Forest Classifier

    7. Random Forest Regressor

    8. Support Vector Classifier

    9. XGBoost Classifier

    10. XGBoost Regressor

  9. Deep Learning Algorithms

    1. Dense Neural Networks

    2. Recurrent Neural Networks

  10. Computer Vision

    1. Convolutional Recurrent Networks

    2. Face detector

    3. Fully Convolutional Neural Networks

    4. Image IO

    5. Image matcher

    6. Yolo

  11. Natural Language Processing

    1. Text Cleaner

    2. Text Vectorizer

  12. Time Series Processing

    1. TS Features selector

  13. TensorFlow2 API

    1. Dense Layer

    2. LSTM Layer

  14. Helpers

    1. Data/Object Logger

    2. Object Selector (5ports)

  15. Testing

    1. Mock module for test

    2. Simple Generator

    3. Simple Receptor