This data was extracted from the 1994 Census bureau database by Ronny Kohavi and Barry Becker (Data Mining and Visualization, Silicon Graphics). A set of reasonably clean records was extracted using the following conditions: ((AAGE>16) && (AGI>100) && (AFNLWGT>1) && (HRSWK>0)).
The prediction task is to determine whether a person makes over $50K a year.
The Adult's Income dataset is collected from kaggle using this https://www.kaggle.com/uciml/adult-census-income.
This dataset has 48842 rows × 15 columns.
- age: continuous.
- workclass: Private, Self-emp-not-inc, Self-emp-inc, Federal-gov, Local-gov, State-gov, Without-pay, Never-worked.
- fnlwgt: continuous.
- education: Bachelors, Some-college, 11th, HS-grad, Prof-school, Assoc-acdm, Assoc-voc, 9th, 7th-8th, 12th, Masters, 1st-4th, 10th, Doctorate, 5th-6th, Preschool.
- education-num: continuous.
- marital-status: Married-civ-spouse, Divorced, Never-married, Separated, Widowed, Married-spouse-absent, Married-AF-spouse.
- occupation: Tech-support, Craft-repair, Other-service, Sales, Exec-managerial, Prof-specialty, Handlers-cleaners, Machine-op-inspct, Adm-clerical, Farming-fishing, Transport-moving, Priv-house-serv, Protective-serv, Armed-Forces.
- relationship: Wife, Own-child, Husband, Not-in-family, Other-relative, Unmarried.
- race: White, Asian-Pac-Islander, Amer-Indian-Eskimo, Other, Black.
- sex: Female, Male.
- capital-gain: continuous.
- capital-loss: continuous.
- hours-per-week: continuous.
- native-country: United-States, Cambodia, England, Puerto-Rico, Canada, Germany, Outlying-US(Guam-USVI-etc), India, Japan, Greece, South, China, Cuba, Iran, Honduras, Philippines, Italy, Poland, Jamaica, Vietnam, Mexico, Portugal, Ireland, France, Dominican-Republic, Laos, Ecuador, Taiwan, Haiti, Columbia, Hungary, Guatemala, Nicaragua, Scotland, Thailand, Yugoslavia, El-Salvador, Trinadad&Tobago, Peru, Hong, Holand-Netherlands.
In this project, we have employed the following steps:
- Data Collection
- Data Cleaning
- Data Analysis
- Data Preprocessing(Preparation)
- Model Training
- Performance Evaluation
We have implemented the following models:
- SVM
- Logistic Regression
- Naive Bayes
We have calculated the performance by calculating Precision, Recall , True positive/negative for all the models and built a Confusion Matrix for the same.
- Gender Distribution based on Income
- Education Count
- Age Vs Marital Status
Here is the tabular view of Accuracy and Time taken to run the model -
| Model | Accuracy | Time |
|---|---|---|
| Logistic Reg - Oversampled | 83.2% | 824.2s |
| Logistic Reg - Undersampled | 82.5% | 102.9s |
| Logistic Reg - SMOTE | 83.7% | 848s |
| SVM - Oversampled | 83% | 303.9s |
| SVM - Undersampled | 81% | 22.89s |
| SVM - SMOTE | 84% | 300.4s |
| Naive Bayes - Oversampled | 76% | 23.41s |
| Naive Bayes - Undersampled | 76% | 3.69s |
| Naive Bayes - SMOTE | 85% | 23.67s |
- Only a few of the traits are closely associated to income projection, according to the correlation table.
- The choice of such columns has a big impact on the model's accuracy. Random Forest is also the best fit for this dataset. Males have a higher rate of incorrect predictions than females.
- The dataset does not have to be perfect or small all of the time. As a result, we must define a method, a method that can be applied to every dataset.
- Here is a step-by-step explanation of the workflow, which you may apply to your own dataset.
- Also, different types of visualizations effectively portray data based on a combination of attribute types (numerical, categorical) and a classification technique that is totally reliant on the dataset.