This is an software application that performs test equalization.
The application can quantify the respondents' ability by showing their predictive count of correctly answered questions.
Note: The following are the demo version and have altered layout and data for demonstration.
Prepare the training data for the model with an intuitive File Explorer interface.
Upon clicking the execution button, it proceeds with two primary functions:
- Training: The IRT (Item Response Theory) model undergoes training
- Redirection: Users are automatically redirected to the report page designed for visualization and search table.
Offer an interactive graph that dynamically illustrates the relationship between the number of correctly answered questions and the abilities of respondents.
- Sliders: Users can manipulate sliders to observe variations in these metrics.
- Visualization: The graph responds instantly to slider adjustments, providing real-time insights into the data.
Designed for conducting real-time searches with a specific focus:
- Ability Level Specification: Users can specify an ability level as a search parameter.
- Targeted Results: The search predicted the count of correctly answered questions associated with the specified ability level.
-
📈 IRT Model Training:
- Developed and integrated a real-time Item Response Theory (IRT) model in the UI.
- Utilized Bayesian estimation techniques for dynamic model adaptation.
- Predicted the count of correctly answered questions given a student's ability.
-
🌐 UI Design:
- Created a user-friendly interface that significantly streamlined statistical analysis.
- Facilitated immediate feedback and specific search.
- 📘 Girth: Item Response Theory (IRT) model analysis. Girth on PyPI
- 🖥️ PyQt5: Creation of the user interface. PyQt5 Information