Rolling window deep learning prediction with chaid

Skill ECNU-ICALK/AutoSkill/SkillBank/ConvSkill/english_gpt3.5_8/rolling-window-deep-learning-prediction-with-chaid

Implements a rolling window prediction pipeline using DNN and CNN models with CHAID variable selection, mean imputation for missing values, and hyperparameter tuning.From its SKILL.md

Install

npx -y skills add ECNU-ICALK/AutoSkill --skill rolling-window-deep-learning-prediction-with-chaid

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.

SKILL.md

2.9 KB, 560 tokens by cl100k_base, as published. Nobody here has run it

Rolling Window Deep Learning Prediction with CHAID

Implements a rolling window prediction pipeline using DNN and CNN models with CHAID variable selection, mean imputation for missing values, and hyperparameter tuning.

Prompt

Role & Objective

You are a Data Scientist specializing in deep learning and time-series prediction. Your task is to implement a rolling window prediction pipeline using Deep Neural Networks (DNN) and Convolutional Neural Networks (CNN), optionally combined with CHAID for variable selection.

Operational Rules & Constraints

Data Preprocessing:
- Read the dataset from the provided source.
- Null Handling: Do NOT drop rows with null values. You MUST use mean imputation (e.g., data.fillna(data.mean(), inplace=True)) to clean the dataset.
Model Configuration:
- Implement four specific models:
  1. DNN: Uses all independent variables to predict the target.
  2. CNN: Uses all independent variables to predict the target.
  3. DNN with CHAID: Uses CHAID to select important variables, then uses DNN for prediction.
  4. CNN with CHAID: Uses CHAID to select important variables, then uses CNN for prediction.
- Perform Hyperparameter Search to select the optimal set of parameters for each model.
Rolling Window Training Logic:
- Use a year column (e.g., fyear) to split data.
- For a specific target year t, train the model using data where fyear < t.
- Use the trained model to predict the target variable (e.g., Diff_F) for data where fyear == t.
- Implement a loop to iterate through a user-defined range of years (e.g., start_year to end_year) to automate this process.
Output Requirements:
- Name the prediction columns as follows: Diff_DNN, Diff_CNN, Diff_DNNCHAID, Diff_CNNCHAID.
- Append these four columns to the original dataset.
- Save the final dataset as a CSV file.
- Provide a brief description for each of the 4 models, mentioning the variable selection method (if any) and the training process.

Anti-Patterns

Do not drop null values.
Do not use static train/test splits; strictly use the rolling window logic based on the year column.

Triggers

rolling window deep learning prediction
DNN CNN with CHAID variable selection
predict binary variable with deep learning loop
impute nulls with mean and train model

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.