Dat science using weka or other toolkit
Budget: ₹600 – ₹1,500 INR
Analytics Task
You will choose, for this mini project, any *one* of the following tasks:
• Classification: The task of building a classifier (using an appropriate train/test split of the data) to predict whether an individual earns <=50k$ or >50k$. Popular classification algorithms include AdaBoost, Decision Trees, Random Forest, SVM, or neural methods.
• Clustering: The task of grouping the individuals in the dataset into a specified number of clusters using an appropriate clustering algorithm. Popular clustering algorithms include K-Means, Hierarchical Agglomerative Clustering, DBSCAN etc.
Once you have decided on one of the above tasks, you will also need to identify an implementation of a classification or clustering. You are required to choose any one technique and are *not* expected to try multiple classifiers or multiple clustering algorithms. You could use implementations of classifiers and clustering algorithms in the Weka toolkit, or within other Python-based machine learning frameworks. When there are hyperparameters (e.g., fraction to be used as training data to train the classifier, number of clusters in the output for clustering), you are encouraged to experiment with several settings of hyperparameters and choose those that could be considered most appropriate.
You will choose, for this mini project, any *one* of the following tasks:
• Classification: The task of building a classifier (using an appropriate train/test split of the data) to predict whether an individual earns <=50k$ or >50k$. Popular classification algorithms include AdaBoost, Decision Trees, Random Forest, SVM, or neural methods.
• Clustering: The task of grouping the individuals in the dataset into a specified number of clusters using an appropriate clustering algorithm. Popular clustering algorithms include K-Means, Hierarchical Agglomerative Clustering, DBSCAN etc.
Once you have decided on one of the above tasks, you will also need to identify an implementation of a classification or clustering. You are required to choose any one technique and are *not* expected to try multiple classifiers or multiple clustering algorithms. You could use implementations of classifiers and clustering algorithms in the Weka toolkit, or within other Python-based machine learning frameworks. When there are hyperparameters (e.g., fraction to be used as training data to train the classifier, number of clusters in the output for clustering), you are encouraged to experiment with several settings of hyperparameters and choose those that could be considered most appropriate.