project_config.yml

# project_config.yml

# Databricks catalog and schema names for saving tables
catalog_name: "mlops_students"
schema_name: "javedhassi"

# Parameters for model training
# Set up Random Forest parameters
random_forest_parameters:
  n_estimators: 200         # Number of trees in the forest
  max_depth: 10             # Maximum depth of each tree
  min_samples_split: 5      # Minimum number of samples required to split a node
  min_samples_leaf: 2       # Minimum number of samples required at each leaf node
  class_weight: balanced    # Automatically balance classes
  random_state: 42

# Set up LightGBM parameters
lgb_parameters:
  learning_rate: 0.03884249148676395
  max_depth: 12
  num_leaves: 413
  min_data_in_leaf: 14
  feature_fraction: 0.7987976913702801
  bagging_fraction: 0.7602261703576205
  bagging_freq: 2
  lambda_l1: 4.735462555910575
  lambda_l2: 4.735028557007343e-06
  n_estimators: 100
  num_class: 5
  metric: multi_logloss
  objective: multiclass
  boosting_type: gbdt

pipeline_steps:
  feature_engineering: true
  scaling: true
  handle_outliers: true
  handle_missing_values: true

logging:
  level: INFO
  file: logs/data_processor.log


# Numerical features
num_features:
  - Basic_Demos-Age    # Age of the individual.
  - CGAS-CGAS_Score    # Clinical Global Assessment Scale (CGAS) Score
  - Physical-BMI       # Body Mass Index (BMI)
  - Physical-Height    # Height of the individual.
  - Physical-Weight    # Weight of the individual.
  - Physical-Waist_Circumference # Waist circumference of the individual.
  - Physical-Diastolic_BP # Diastolic blood pressure of the individual.
  - Physical-HeartRate    # Heart rate of the individual.
  - Physical-Systolic_BP  # Systolic blood pressure of the individual.
  - Fitness_Endurance-Max_Stage # Max stage of endurance test.
  - Fitness_Endurance-Time_Mins # Time taken to complete the endurance test in minutes.
  - Fitness_Endurance-Time_Sec # Time taken to complete the endurance test in seconds.
  - FGC-FGC_CU # Functional Games Characteristics (FGC) FGC-CU
  - FGC-FGC_CU_Zone # Functional Games Characteristics (FGC) FGC-CU Zone
  - FGC-FGC_GSND    # Functional Games Characteristics (FGC) FGC-GSND
  - FGC-FGC_GSND_Zone # Functional Games Characteristics (FGC) FGC-GSND Zone
  - FGC-FGC_GSD  # Functional Games Characteristics (FGC) FGC-GSD
  - FGC-FGC_GSD_Zone  # Functional Games Characteristics (FGC) FGC-GSD Zone
  - FGC-FGC_PU  # Functional Games Characteristics (FGC) FGC-PU
  - FGC-FGC_PU_Zone  # Functional Games Characteristics (FGC) FGC-PU Zone
  - FGC-FGC_SRL  # Functional Games Characteristics (FGC) FGC-SRL
  - FGC-FGC_SRL_Zone  # Functional Games Characteristics (FGC) FGC-SRL Zone
  - FGC-FGC_SRR  # Functional Games Characteristics (FGC) FGC-SRR
  - FGC-FGC_SRR_Zone  # Functional Games Characteristics (FGC) FGC-SRR Zone
  - FGC-FGC_TL  # Functional Games Characteristics (FGC) FGC-TL
  - FGC-FGC_TL_Zone   # Functional Games Characteristics (FGC) FGC-TL Zone
  - BIA-BIA_Activity_Level_num # Bioelectrical Impedance Analysis (BIA)
  - BIA-BIA_BMC  # BIA-BMC
  - BIA-BIA_BMI  # BIA-BMI
  - BIA-BIA_BMR  # BIA-BMR
  - BIA-BIA_DEE   # BIA-DEE
  - BIA-BIA_ECW  # BIA-ECW
  - BIA-BIA_FFM  # BIA-FFM
  - BIA-BIA_FFMI # BIA-FFMI
  - BIA-BIA_FMI # BIA-FMI
  - BIA-BIA_Fat  # BIA-Fat
  - BIA-BIA_Frame_num  # BIA-Frame
  - BIA-BIA_ICW  # BIA-ICW
  - BIA-BIA_LDM  # BIA-LDM
  - BIA-BIA_LST  # BIA-LST
  - BIA-BIA_SMM  # BIA-SMM
  - BIA-BIA_TBW  # BIA-TBW
  - PAQ_A-PAQ_A_Total # Physical Activity Questionnaire (PAQ) A Total
  - PAQ_C-PAQ_C_Total # Physical Activity Questionnaire (PAQ) C Total
  - PCIAT-PCIAT_01  # Parent-Child Internet Addiction Test
  - PCIAT-PCIAT_02
  - PCIAT-PCIAT_03
  - PCIAT-PCIAT_04
  - PCIAT-PCIAT_05
  - PCIAT-PCIAT_06
  - PCIAT-PCIAT_07
  - PCIAT-PCIAT_08
  - PCIAT-PCIAT_09
  - PCIAT-PCIAT_10
  - PCIAT-PCIAT_11
  - PCIAT-PCIAT_12
  - PCIAT-PCIAT_13
  - PCIAT-PCIAT_14
  - PCIAT-PCIAT_15
  - PCIAT-PCIAT_16
  - PCIAT-PCIAT_17
  - PCIAT-PCIAT_18
  - PCIAT-PCIAT_19
  - PCIAT-PCIAT_20
  - PCIAT-PCIAT_Total
  - SDS-SDS_Total_Raw  # Sleep Disturbance Scale (SDS) Total Raw
  - SDS-SDS_Total_T    # Sleep Disturbance Scale (SDS) Total T
  # - PreInt_EduHx-Season
  - PreInt_EduHx-computerinternet_hoursday

# Categorical features
cat_features:
  - Basic_Demos-Enroll_Season
  - Basic_Demos-Sex    # Gender of the individual.
  - CGAS-Season       # Clinical Global Assessment Scale (CGAS) Season
  - Physical-Season   # season during which the physical measurements were taken
  - Fitness_Endurance-Season # Season for fitness recording
  - FGC-Season        # Functional Games Characteristics (FGC) Season
  - BIA-Season        # Bioelectrical Impedance Analysis (BIA) Season
  - PAQ_A-Season      # Physical Activity Questionnaire (PAQ) A Season
  - PAQ_C-Season      # Physical Activity Questionnaire (PAQ) C Season
  - PCIAT-Season      # PCIAT-Season
  - SDS-Season      # Sleep Disturbance Scale (SDS) Season

# Target column
target: sii