danfo.OneHotEncoder
Encode categorical features as a one-hot numeric array.
class danfo.OneHotEncoder
danfo.js provides the OneHotEncoder class for encoding values in Series and Arrays to one-hot numeric arrays. This is mostly used as a preprocessing step before most machine learning tasks.
Examples
Convert Series to Dummy codes
Node
1
const dfd = require("danfojs-node")
2
3
let data = {
4
fruits: ['pear', 'mango', "pawpaw", "mango", "bean"],
5
Count: [20, 30, 89, 12, 30],
6
Country: ["NG", "NG", "GH", "RU", "RU"]
7
}
8
9
10
let df = new dfd.DataFrame(data)
11
let encode = new dfd.OneHotEncoder()
12
13
encode.fit(df['fruits'])
14
console.log(encode);
15
16
let sf_enc = encode.transform(df['fruits'].values)
17
console.log(sf_enc)
18
19
let new_sf = encode.transform(["mango", "bean"])
20
console.log(new_sf)
Copied!
Output
1
OneHotEncoder { '$labels': [ 'pear', 'mango', 'pawpaw', 'bean' ] }
2
[
3
[ 1, 0, 0, 0 ],
4
[ 0, 1, 0, 0 ],
5
[ 0, 0, 1, 0 ],
6
[ 0, 1, 0, 0 ],
7
[ 0, 0, 0, 1 ]
8
]
9
[ [ 0, 1, 0, 0 ], [ 0, 0, 0, 1 ] ]
Copied!
Labels not found in the original data used for fitting are represented with 0s all through
Last modified 5mo ago
Copy link