asdfaqer.github.io/Homepage.html at main · asdfaqer/asdfaqer.github.io · GitHub

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
154
155
156
157
158
159
160
161
162
163
164

<!DOCTYPE html>
<html>
<head>
    <title>NN Classification Demo</title>
</head>
<body>
    <h1>NN Classification Demo</h1>
    <p>Below is a demonstration of neural network classification in action.</p>

    <button type="button">Distribution 1</button>
    <button type="button">Distribution 2</button>
    <button type="button">Distribution 3</button>

    <button onclick="classify()">Classify</button>

    <select id="mySelect">
        <option value="line of best fit">line of best fit</option>
        <option value="relu classification">relu classification</option>
        <option value="tanh classification">tanh classification</option>
    </select>

    <div id="my-div">
        <canvas id="myCanvas" width="500" height="500" style="border:1px solid #000000;">
            Your browser does not support the HTML5 canvas element.
        </canvas>

        <script>
            const CANVAS_WIDTH = 500;
            const CANVAS_HEIGHT = 500;
            var c = document.getElementById("myCanvas");
            var ctx = c.getContext("2d");

            function draw_point(x, y, color) {
                ctx.fillStyle = color;
                ctx.fillRect(x, y, 3, 3);
            }

            function draw_line(x1, y1, x2, y2, color) {
                ctx.strokeStyle = color;
                ctx.beginPath();
                ctx.moveTo(x1, y1);
                ctx.lineTo(x2, y2);
                ctx.stroke();
            }

            function draw_points_list(points_list) {
                for (let i = 0; i < points_list.length; i++) {
                    draw_point(points_list[i][0], points_list[i][1], points_list[i][2]);
                }
            }

            // --- NEW: Function to draw the light-colored background ---
            function drawBackground(slope, y_int) {
                const lightRed = 'rgb(255, 220, 220)';
                const lightGreen = 'rgb(220, 255, 220)';
                const step = 5; // Draw a 5x5 pixel square for performance

                for (let x = 0; x < CANVAS_WIDTH; x += step) {
                    for (let y = 0; y < CANVAS_HEIGHT; y += step) {
                        const prediction_is_red = (y < slope * x + y_int);
                        ctx.fillStyle = prediction_is_red ? lightRed : lightGreen;
                        ctx.fillRect(x, y, step, step);
                    }
                }
            }

            const points = [[61, 436, 'rgb(0, 150, 0)'], [130, 337, 'rgb(0, 150, 0)'], [66, 329, 'rgb(0, 150, 0)'], [83, 475, 'rgb(0, 150, 0)'], [290, 495, 'rgb(0, 150, 0)'], [134, 294, 'rgb(0, 150, 0)'], [285, 432, 'rgb(0, 150, 0)'], [194, 137, 'rgb(150, 0, 0)'], [185, 257, 'rgb(0, 150, 0)'], [412, 349, 'rgb(150, 0, 0)'], [50, 479, 'rgb(0, 150, 0)'], [77, 373, 'rgb(0, 150, 0)'], [283, 98, 'rgb(150, 0, 0)'], [426, 42, 'rgb(150, 0, 0)'], [443, 225, 'rgb(150, 0, 0)'], [345, 29, 'rgb(150, 0, 0)'], [421, 374, 'rgb(150, 0, 0)'], [199, 81, 'rgb(150, 0, 0)'], [52, 494, 'rgb(0, 150, 0)'], [214, 248, 'rgb(0, 150, 0)'], [72, 389, 'rgb(0, 150, 0)'], [487, 59, 'rgb(150, 0, 0)'], [274, 473, 'rgb(0, 150, 0)'], [103, 91, 'rgb(150, 0, 0)'], [139, 240, 'rgb(0, 150, 0)'], [444, 131, 'rgb(150, 0, 0)'], [240, 389, 'rgb(0, 150, 0)'], [81, 293, 'rgb(0, 150, 0)'], [78, 189, 'rgb(0, 150, 0)'], [22, 182, 'rgb(0, 150, 0)'], [292, 119, 'rgb(150, 0, 0)'], [289, 161, 'rgb(150, 0, 0)'], [352, 260, 'rgb(150, 0, 0)'], [215, 62, 'rgb(150, 0, 0)'], [50, 125, 'rgb(0, 150, 0)'], [493, 473, 'rgb(150, 0, 0)'], [142, 155, 'rgb(0, 150, 0)'], [27, 310, 'rgb(0, 150, 0)'], [265, 251, 'rgb(150, 0, 0)'], [52, 166, 'rgb(0, 150, 0)'], [408, 73, 'rgb(150, 0, 0)'], [130, 71, 'rgb(150, 0, 0)'], [270, 21, 'rgb(150, 0, 0)'], [53, 191, 'rgb(0, 150, 0)'], [177, 185, 'rgb(0, 150, 0)'], [383, 164, 'rgb(150, 0, 0)'], [188, 349, 'rgb(0, 150, 0)'], [226, 62, 'rgb(150, 0, 0)'], [397, 226, 'rgb(150, 0, 0)'], [300, 451, 'rgb(0, 150, 0)'], [308, 221, 'rgb(150, 0, 0)'], [296, 157, 'rgb(150, 0, 0)'], [345, 251, 'rgb(150, 0, 0)'], [207, 429, 'rgb(0, 150, 0)'], [423, 42, 'rgb(150, 0, 0)'], [370, 402, 'rgb(0, 150, 0)'], [201, 461, 'rgb(0, 150, 0)'], [160, 437, 'rgb(0, 150, 0)'], [259, 17, 'rgb(150, 0, 0)'], [218, 125, 'rgb(150, 0, 0)'], [9, 179, 'rgb(0, 150, 0)'], [372, 375, 'rgb(0, 150, 0)'], [322, 297, 'rgb(150, 0, 0)'], [497, 221, 'rgb(150, 0, 0)'], [309, 228, 'rgb(150, 0, 0)'], [337, 210, 'rgb(150, 0, 0)'], [500, 214, 'rgb(150, 0, 0)'], [419, 39, 'rgb(150, 0, 0)'], [367, 71, 'rgb(150, 0, 0)'], [350, 368, 'rgb(0, 150, 0)'], [355, 464, 'rgb(0, 150, 0)'], [336, 80, 'rgb(150, 0, 0)'], [490, 63, 'rgb(150, 0, 0)'], [451, 264, 'rgb(150, 0, 0)'], [470, 52, 'rgb(150, 0, 0)'], [299, 319, 'rgb(0, 150, 0)'], [294, 427, 'rgb(0, 150, 0)'], [43, 403, 'rgb(0, 150, 0)'], [329, 236, 'rgb(150, 0, 0)'], [453, 213, 'rgb(150, 0, 0)'], [422, 150, 'rgb(150, 0, 0)'], [438, 20, 'rgb(150, 0, 0)'], [238, 447, 'rgb(0, 150, 0)'], [290, 111, 'rgb(150, 0, 0)'], [376, 297, 'rgb(150, 0, 0)'], [247, 396, 'rgb(0, 150, 0)'], [234, 492, 'rgb(0, 150, 0)'], [132, 305, 'rgb(0, 150, 0)'], [423, 106, 'rgb(150, 0, 0)'], [102, 282, 'rgb(0, 150, 0)'], [360, 434, 'rgb(0, 150, 0)'], [53, 191, 'rgb(0, 150, 0)'], [58, 420, 'rgb(0, 150, 0)'], [443, 402, 'rgb(150, 0, 0)'], [164, 454, 'rgb(0, 150, 0)'], [81, 454, 'rgb(0, 150, 0)'], [366, 110, 'rgb(150, 0, 0)'], [265, 438, 'rgb(0, 150, 0)'], [172, 498, 'rgb(0, 150, 0)'], [251, 65, 'rgb(150, 0, 0)'], [431, 315, 'rgb(150, 0, 0)'], [109, 142, 'rgb(0, 150, 0)'], [240, 382, 'rgb(0, 150, 0)'], [496, 24, 'rgb(150, 0, 0)'], [209, 138, 'rgb(150, 0, 0)'], [160, 130, 'rgb(150, 0, 0)'], [328, 102, 'rgb(150, 0, 0)'], [223, 431, 'rgb(0, 150, 0)'], [287, 36, 'rgb(150, 0, 0)'], [166, 421, 'rgb(0, 150, 0)'], [164, 395, 'rgb(0, 150, 0)'], [150, 226, 'rgb(0, 150, 0)'], [152, 435, 'rgb(0, 150, 0)'], [347, 487, 'rgb(0, 150, 0)'], [262, 37, 'rgb(150, 0, 0)'], [464, 26, 'rgb(150, 0, 0)'], [460, 256, 'rgb(150, 0, 0)'], [349, 124, 'rgb(150, 0, 0)'], [135, 319, 'rgb(0, 150, 0)'], [97, 268, 'rgb(0, 150, 0)'], [374, 301, 'rgb(150, 0, 0)'], [195, 285, 'rgb(0, 150, 0)'], [373, 186, 'rgb(150, 0, 0)'], [106, 366, 'rgb(0, 150, 0)'], [301, 251, 'rgb(150, 0, 0)'], [236, 177, 'rgb(150, 0, 0)'], [64, 91, 'rgb(0, 150, 0)'], [217, 80, 'rgb(150, 0, 0)'], [4, 311, 'rgb(0, 150, 0)'], [22, 354, 'rgb(0, 150, 0)'], [470, 275, 'rgb(150, 0, 0)'], [392, 310, 'rgb(150, 0, 0)'], [64, 10, 'rgb(150, 0, 0)'], [487, 92, 'rgb(150, 0, 0)'], [285, 308, 'rgb(0, 150, 0)'], [436, 425, 'rgb(150, 0, 0)'], [132, 97, 'rgb(150, 0, 0)'], [128, 451, 'rgb(0, 150, 0)'], [110, 373, 'rgb(0, 150, 0)'], [49, 66, 'rgb(0, 150, 0)'], [332, 32, 'rgb(150, 0, 0)'], [43, 65, 'rgb(0, 150, 0)'], [353, 75, 'rgb(150, 0, 0)'], [333, 140, 'rgb(150, 0, 0)'], [348, 375, 'rgb(0, 150, 0)'], [151, 70, 'rgb(150, 0, 0)'], [116, 399, 'rgb(0, 150, 0)'], [66, 30, 'rgb(150, 0, 0)'], [66, 421, 'rgb(0, 150, 0)'], [280, 104, 'rgb(150, 0, 0)']];
            // Draw the initial points
            for (let inx = 0; inx < points.length; inx++) {
                draw_point(points[inx][0], points[inx][1], points[inx][2]);
            }
        </script>
    </div>

    <script>
        // --- State variables for our animation loop ---
        let best_slope, best_y_int, best_score, currentIteration;
        let animationFrameId = null; // To keep track of the animation request
        const maxIterations = 1000;
        const red_label = 'rgb(150, 0, 0)';

        // This function will be called to start or restart the process
        function startFindingBestFit() {

            // If an animation is already running, cancel it before starting a new one
            if (animationFrameId) {
                cancelAnimationFrame(animationFrameId);
            }

            // --- Reset the state for a new search ---
            best_slope = -1;
            best_y_int = 450;
            best_score = 0;
            currentIteration = 0;

            // Clear the canvas and redraw the points to start fresh
            ctx.clearRect(0, 0, c.width, c.height);
            draw_points_list(points);

            // Kick off the animation loop
            findBestFitStep();
        }

        // --- This function performs ONE STEP of the search ---
        function findBestFitStep() {
            // 1. Check the stop condition
            if (currentIteration >= maxIterations) {
                console.log("Search complete. Final score:", best_score);
                return; // Stop the animation
            }

            // 2. Perform one iteration of the search logic
            let guess_slope = Math.tan(Math.atan(best_slope) + (Math.random() - 0.5) * 360 / (1 + Math.log(currentIteration + 1)));
            let guess_y_int = best_y_int + (Math.random() - 0.5) * 1000 / (1 + Math.log(currentIteration + 1));
            let current_score = 0;

            for (let i = 0; i < points.length; i++) {
                const p = points[i];
                const prediction_is_red = (p[1] < guess_slope * p[0] + guess_y_int);
                if (prediction_is_red === (p[2] === red_label)) {
                    current_score++;
                }
            }

            // 3. If we found a new best, UPDATE and REDRAW
            if (current_score > best_score) {
                best_score = current_score;
                best_slope = guess_slope;
                best_y_int = guess_y_int;

                console.log(`New best at iteration ${currentIteration}: score ${best_score}`);

                // --- MODIFIED: Redraw the canvas with background, points, and line ---
                // 1. Draw the new light-colored background based on the classifier
                drawBackground(best_slope, best_y_int);

                // 2. Draw the original, darker points on top of the background
                draw_points_list(points);

                // 3. Draw the new best-fit line on top of everything
                draw_line(0, best_y_int, CANVAS_WIDTH, best_y_int + CANVAS_WIDTH * best_slope, 'blue');
            }

            // 4. Schedule the next step
            currentIteration++;
            animationFrameId = requestAnimationFrame(findBestFitStep);
        }

        function classify() {
            const mySelect = document.getElementById('mySelect');
            if (mySelect.value !== 'line of best fit') {
                alert("This classifier is not implemented yet.");
                return;
            }
            else if (mySelect.value === 'line of best fit') {
                startFindingBestFit();
                // Note: The rest of this function will run before startFindingBestFit finishes.
                // The animation loop is the part that visually updates the canvas.
            }
        }
    </script>
</body>
</html>