stock-trek 0.4.3

Stock Trek time-series analysis
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

use crate::statistics::decompose;

#[derive(Clone, Default)]
pub struct Decompose;

impl Decompose {
    pub fn loess_smooth(&self, values: &[f64], span: usize) -> Vec<f64> {
        decompose::loess_smooth(values, span)
    }
    pub fn seasonal_decompose(
        &self,
        time_series_values: &[f64],
        seasonal_period_length: usize,
    ) -> (Vec<f64>, Vec<f64>, Vec<f64>) {
        decompose::seasonal_decompose(time_series_values, seasonal_period_length)
    }
    pub fn seasonal_trend_decomposition_using_loess(
        &self,
        time_series_values: &[f64],
        seasonal_period_length: usize,
    ) -> (Vec<f64>, Vec<f64>, Vec<f64>) {
        decompose::seasonal_trend_decomposition_using_loess(
            time_series_values,
            seasonal_period_length,
        )
    }
}

pub fn seasonal_decompose(
    time_series_values: &[f64],
    seasonal_period_length: usize,
) -> (Vec<f64>, Vec<f64>, Vec<f64>) {
    let n = time_series_values.len();
    if n == 0 || seasonal_period_length == 0 || seasonal_period_length > n {
        return (vec![], vec![], vec![]);
    }
    // Trend via moving average
    let mut trend = vec![0.0; n];
    let half = seasonal_period_length / 2;
    for (i, trend_item) in trend.iter_mut().enumerate() {
        let start = i.saturating_sub(half);
        let end = usize::min(n, i + half + 1);
        let window = &time_series_values[start..end];
        *trend_item = window.iter().sum::<f64>() / window.len() as f64;
    }
    // Detrended
    let detrended: Vec<f64> = time_series_values
        .iter()
        .zip(trend.iter())
        .map(|(x, t)| x - t)
        .collect();
    // Seasonal component
    let mut seasonal = vec![0.0; n];
    let mut seasonal_means = vec![0.0; seasonal_period_length];
    let mut counts = vec![0usize; seasonal_period_length];
    for (i, detrended_item) in detrended.iter().enumerate() {
        let idx = i % seasonal_period_length;
        seasonal_means[idx] += detrended_item;
        counts[idx] += 1;
    }
    for i in 0..seasonal_period_length {
        if counts[i] > 0 {
            seasonal_means[i] /= counts[i] as f64;
        }
    }
    for i in 0..n {
        seasonal[i] = seasonal_means[i % seasonal_period_length];
    }
    // Residual
    let residual: Vec<f64> = time_series_values
        .iter()
        .zip(trend.iter())
        .zip(seasonal.iter())
        .map(|((x, t), s)| x - t - s)
        .collect();
    (trend, seasonal, residual)
}

pub fn seasonal_trend_decomposition_using_loess(
    time_series_values: &[f64],
    seasonal_period_length: usize,
) -> (Vec<f64>, Vec<f64>, Vec<f64>) {
    let n = time_series_values.len();
    if n == 0 || seasonal_period_length == 0 || seasonal_period_length > n {
        return (vec![], vec![], vec![]);
    }
    // Trend via LOESS
    let trend = decompose::loess_smooth(time_series_values, seasonal_period_length);
    // Detrended
    let detrended: Vec<f64> = time_series_values
        .iter()
        .zip(trend.iter())
        .map(|(x, t)| x - t)
        .collect();
    // Seasonal component (same approach as classical)
    let mut seasonal = vec![0.0; n];
    let mut seasonal_means = vec![0.0; seasonal_period_length];
    let mut counts = vec![0usize; seasonal_period_length];
    for (i, detrended_item) in detrended.iter().enumerate() {
        let idx = i % seasonal_period_length;
        seasonal_means[idx] += detrended_item;
        counts[idx] += 1;
    }
    for i in 0..seasonal_period_length {
        if counts[i] > 0 {
            seasonal_means[i] /= counts[i] as f64;
        }
    }
    for i in 0..n {
        seasonal[i] = seasonal_means[i % seasonal_period_length];
    }
    // Residual
    let residual: Vec<f64> = time_series_values
        .iter()
        .zip(trend.iter())
        .zip(seasonal.iter())
        .map(|((x, t), s)| x - t - s)
        .collect();
    (trend, seasonal, residual)
}

pub fn loess_smooth(values: &[f64], span: usize) -> Vec<f64> {
    let n = values.len();
    let mut result = vec![0.0; n];
    for i in 0..n {
        let start = i.saturating_sub(span);
        let end = usize::min(n, i + span + 1);
        let mut weighted_sum = 0.0;
        let mut weight_total = 0.0;
        for (j, j_value) in values.iter().enumerate().take(end).skip(start) {
            let dist = (i as isize - j as isize).abs() as f64;
            let w = (1.0 - (dist / span as f64).powi(3)).powi(3).max(0.0);
            weighted_sum += w * j_value;
            weight_total += w;
        }
        result[i] = if weight_total > 0.0 {
            weighted_sum / weight_total
        } else {
            values[i]
        };
    }
    result
}