rust_nb 0.1.1

A simple but generic Naive Bayes Model in Rust.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153

#[cfg(test)]
mod rust_nb {
    use rust_nb::{Feature, FeatureType, Model};

    #[test]
    fn model_works_simple_case() {
        let mut model = Model::new();

        let input_train = vec![
            (
                "happy".to_owned(),
                vec![Feature {
                    feature_type: FeatureType::Text,
                    name: "my_words".to_owned(),
                    value: "good".to_owned(),
                }],
            ),
            (
                "sad".to_owned(),
                vec![Feature {
                    feature_type: FeatureType::Text,
                    name: "my_words".to_owned(),
                    value: "bad".to_owned(),
                }],
            ),
        ];
        model.train("test_model", &input_train);

        let input_test = vec![Feature {
            feature_type: FeatureType::Text,
            name: "my_words".to_owned(),
            value: "no opinion".to_owned(),
        }];
        let result = model.predict("test_model", &input_test);
        assert_eq!(0.5, *result.get("happy").unwrap());
        assert_eq!(0.5, *result.get("sad").unwrap());

        let input_test = vec![Feature {
            feature_type: FeatureType::Text,
            name: "my_words".to_owned(),
            value: "GOOD".to_owned(),
        }];
        let result = model.predict("test_model", &input_test);
        assert!((0.6666666666666666 - *result.get("happy").unwrap()).abs() < 1e-10);
        assert!((0.3333333333333333 - *result.get("sad").unwrap()).abs() < 1e-10);
    }

    #[test]
    fn model_works() {
        let mut model = Model::new();

        let input_train = vec![
            (
                "happy".to_owned(),
                vec![Feature {
                    feature_type: FeatureType::Text,
                    name: "my_words".to_owned(),
                    value: "The weather is so good".to_owned(),
                }],
            ),
            (
                "sad".to_owned(),
                vec![Feature {
                    feature_type: FeatureType::Text,
                    name: "my_words".to_owned(),
                    value: "that food tastes so bad".to_owned(),
                }],
            ),
        ];
        model.train("test_model", &input_train);

        let input_test = vec![Feature {
            feature_type: FeatureType::Text,
            name: "my_words".to_owned(),
            value: "thinking about the weather ...".to_owned(),
        }];
        let result = model.predict("test_model", &input_test);

        assert!((0.8 - *result.get("happy").unwrap()).abs() < 1e-10);
        assert!((0.2 - *result.get("sad").unwrap()).abs() < 1e-10);
    }

    #[test]
    fn model_gaussian_works() {
        let mut model = Model::new();

        let input_train = vec![
            (
                "eur".to_owned(),
                vec![Feature {
                    feature_type: FeatureType::GaussianStd,
                    name: "age".to_owned(),
                    value: "20".to_owned(),
                }],
            ),
            (
                "eur".to_owned(),
                vec![Feature {
                    feature_type: FeatureType::GaussianStd,
                    name: "age".to_owned(),
                    value: "30".to_owned(),
                }],
            ),
            (
                "usd".to_owned(),
                vec![Feature {
                    feature_type: FeatureType::GaussianStd,
                    name: "age".to_owned(),
                    value: "40".to_owned(),
                }],
            ),
            (
                "usd".to_owned(),
                vec![Feature {
                    feature_type: FeatureType::GaussianStd,
                    name: "age".to_owned(),
                    value: "50".to_owned(),
                }],
            ),
            (
                "eur".to_owned(),
                vec![Feature {
                    feature_type: FeatureType::GaussianStd,
                    name: "age".to_owned(),
                    value: "40".to_owned(),
                }],
            ),
        ];
        model.train("test_model", &input_train);

        let input_test = vec![Feature {
            feature_type: FeatureType::GaussianStd,
            name: "age".to_owned(),
            value: "23".to_owned(),
        }];
        let result = model.predict("test_model", &input_test);

        println!("{:?}", result);

        // let mut p_eur = 0.0312254; // https://www.wolframalpha.com/input/?i=normalpdf(23.0,+30.0,+10.0)
        // let mut p_usd = 0.000446108; // https://www.wolframalpha.com/input/?i=normalpdf(23.0,+45.0,+7.0710678118654755)

        // p_usd = p_usd * (2.0 / 5.0);
        // p_eur = p_eur * (3.0 / 5.0);

        // p_usd = p_usd / (p_usd + p_eur);
        // p_eur = 1.0 - p_usd;

        assert!((0.1673554586478428 - *result.get("usd").unwrap()).abs() < 1e-6);
        assert!((0.8326445413521572 - *result.get("eur").unwrap()).abs() < 1e-6);
    }

}