Which ML algorithm should I use for following use case for classification and Why?


I have data in table format having total 3 columns. One column for label, other two columns are features. So, such 30 rows(1 row contains 2 feature and 1 label) make one set of data with all 30 rows having same. In these 30 rows there could be some patterns of features. There could be more such sets of 30 rows(for 30 rows label will be same). So there will be 6 to 7 labels or class. I want algorithm to predict one label from 30 rows(one set). Which algorithm should I choose and how should I pre-process data?

Mitesh Patel

Posted 2020-04-07T06:01:11.807

Reputation: 149



You can check https://autogluon.mxnet.io/

from autogluon import TabularPrediction as task
predictor = task.fit(train_data=task.Dataset(file_path=TRAIN_DATA.csv), label=COLUMN_NAME)
predictions = predictor.predict(task.Dataset(file_path=TEST_DATA.csv))


Posted 2020-04-07T06:01:11.807

Reputation: 339


For pre-processing, you can use StandardScalar or MinMaxScalar from sklearn.preprocessing library in python. For classification Decision Trees and Random Forests, Neural Networks are good ML algorithms.
Check this.


Posted 2020-04-07T06:01:11.807