Video Showcase

A longer video with audio can be accessed here.

Rice Taste Prediction with Fuzzy Logic System

This MATLAB project implements a Fuzzy Logic System (FIS) to predict the taste of rice based on a given dataset. The FIS is designed to model the taste as the output, utilizing various input variables, such as stickiness.

How to Use

Follow these steps to utilize the Rice Taste Prediction Fuzzy Logic System:

1. Opening the GUI:

   • Run the app.mlapp file to access the graphical user interface (GUI).

2. Evaluating Results:

   • provide inputs in the 4 text fields and click enter (all inputs must be in the correct range or else you will get a warning in the console).

Utilize the provided GUI and scripts to streamline the process of taste prediction and explore the functionalities for your specific use case.

Directory Overview

• app.mlapp: Graphical User Interface (GUI) for the FIS.
• objectiveFunc.m: Helper function to calculate Root Mean Square Error (RMSE) for FIS evaluation.
• riceTaste.m: Scripts for FIS generation and tuning.
• riceTasteFIS.fis: Latest saved FIS file for quick loading and usage.
• readme.md: Documentation file (you are here).

Dataset

The rice taste dataset includes the following input variables:

• Feature 1
• Feature 2
• Feature 3
• Feature 4

And the output variable: Taste score

FIS Data Preprocessing

The code reads the rice taste dataset and separates input and output variables.

Baseline FIS

A baseline fuzzy inference system (FIS) is generated using the genfis function with Fuzzy C-Means clustering and Mamdani type rules.

Rulebase Evaluation

The FIS is evaluated, and its output is plotted using centroid defuzzification.

Tuning FIS

The FIS is further tuned in two steps:

1. Learning rules from scratch using a Genetic Algorithm (GA).
2. Tuning membership functions using a Particle Swarm Optimization (PSO) algorithm.

Evaluation and Comparison

The costs of different FIS configurations are calculated based on the root mean square error (RMSE) using the objectiveFunc function.

Results

The costs of the baseline FIS, FIS with learned rules, and FIS with tuned membership functions are displayed.

Rulebases

The rulebases of the baseline FIS and the FIS with learned rules are shown.

Objective Function

The objective function calculates the RMSE between actual and predicted values.

Feel free to reach out for any questions or improvements!

Contact: zhengjiawen44@gmail.com