七段顯示器辨識 7*4*4*10 | JavaScript 程式專案

{"id":71843,"student_id":10,"content":"var a = createSprite(\"b.png\");\na.moveTo(100, 50);\n\nvar b = createSprite(\"b.png\");\nb.direction = 0;\nb.moveTo(40, 110);\n\nvar c = createSprite(\"b.png\");\nc.direction = 0;\nc.moveTo(160, 110);\n\nvar d = createSprite(\"b.png\");\nd.moveTo(100, 170);\n\nvar e = createSprite(\"b.png\");\ne.direction = 0;\ne.moveTo(40, 230);\n\nvar f = createSprite(\"b.png\");\nf.direction = 0;\nf.moveTo(160, 230);\n\nvar g = createSprite(\"b.png\");\ng.moveTo(100, 290);\n\nvar blocks = [a, b, c, d, e, f, g];\n\nvar data = [\n    [[1, 1, 1, 0, 1, 1, 1], [1,-1,-1,-1,-1,-1,-1,-1,-1,-1]], // 0\n    [[0, 0, 1, 0, 0, 1, 0], [-1,1,-1,-1,-1,-1,-1,-1,-1,-1]], // 1\n    [[1, 0, 1, 1, 1, 0, 1], [-1,-1,1,-1,-1,-1,-1,-1,-1,-1]], // 2\n    [[1, 0, 1, 1, 0, 1, 1], [-1,-1,-1,1,-1,-1,-1,-1,-1,-1]], // 3\n    [[0, 1, 1, 1, 0, 1, 0], [-1,-1,-1,-1,1,-1,-1,-1,-1,-1]], // 4\n    [[1, 1, 0, 1, 0, 1, 1], [-1,-1,-1,-1,-1,1,-1,-1,-1,-1]], // 5\n    [[1, 1, 0, 1, 1, 1, 1], [-1,-1,-1,-1,-1,-1,1,-1,-1,-1]], // 6\n    [[1, 1, 1, 0, 0, 1, 0], [-1,-1,-1,-1,-1,-1,-1,1,-1,-1]], // 7\n    [[1, 1, 1, 1, 1, 1, 1], [-1,-1,-1,-1,-1,-1,-1,-1,1,-1]], // 8\n    [[1, 1, 1, 1, 0, 1, 0], [-1,-1,-1,-1,-1,-1,-1,-1,-1,1]], // 9\n];\n\n\nvar nn = new NeuralNet().init(7, 4, 4, 10);\nnn.train(data, 1000, 0.02, 0.03);\n\n\nblocks.forEach((b) =\u003e {\n    b.on('click', () =\u003e {\n        b.opacity = b.opacity == 1 ? 0.2 : 1;  \n    });\n});\n\n\nforever(function() {\n    var inputs = blocks.map((b) =\u003e {\n        return b.opacity == 1 ? 1: 0;\n    });\n    var output = nn.update(inputs);\n    \n    for (var i=0; i\u003c10; i++) {\n        print(output[i], 400, 200 + 15*i);\n    }\n    print(output.indexOf(Math.max(...output)), 400, 400, 'red');\n});\n\n\n\n","created_at":"2019-06-14T14:35:51.376+08:00","updated_at":"2019-11-11T14:05:51.322+08:00","name":"七段顯示器辨識  7*4*4*10","language":"javascript","screenshot":{"url":"https://cdn0.koding.school/uploads/project/screenshot/71843/2a5fa01b64ef273d98988cfa7a81aabe.jpg"},"parent_id":71802,"plugin":"// noprotect\n\nfunction sigmoid(x) {\n    return (Math.exp(x) - Math.exp(-x)) / (Math.exp(x) + Math.exp(-x));\n}\n\nfunction dsigmoid(x) {\n    return 1.0 - x*x;\n}\n\n\n// 建立大小為 n 的陣列並填入初始值 fill\nvar makeArray = function(n, fill) {\n    var a = [];\n    for (var i = 0; i \u003c n; i++)\n        a.push(fill);\n    return a;\n}\n\n// 建立大小為 I*J 的矩陣並填入初始值 fill\nvar makeMatrix = function(I, J, fill) {\n    var m = [];\n    for (var i = 0; i \u003c I; i++)\n        m.push(makeArray(J, fill));\n    return m;\n}\n\n// numbersToStr():以精確度為 precision 個小數來輸出陣列 array\nvar numbersToStr = function(array, precision) {\n    var rzStr = \"\";\n    for (var i = 0; i \u003c array.length; i++) {\n        if (array[i] \u003e= 0)\n            rzStr += \" \"+array[i].toFixed(precision)+\" \";\n        else\n            rzStr += array[i].toFixed(precision)+\" \";\n    }\n    return rzStr;\n}\n\n// rand():取得 a 到 b 之間的一個隨機亂數\nvar rand = function(a, b) {\n    return (b-a)*Math.random() + a;\n}\n\n// sigmoid(x)=tanh(x)\nfunction sigmoid(x) {\n    var tanh = (Math.exp(x) - Math.exp(-x)) / (Math.exp(x) + Math.exp(-x));\n    return tanh; // 雙曲正切函數\n}\n\n// dsigmoid(x)=1-x^2;\n// 參考：http://pynopticon.googlecode.com/svn/trunk/src/vlfeat/toolbox/special/dsigmoid.m\n// 參考：http://en.wikipedia.org/wiki/Sigmoid_function\nfunction dsigmoid(x) {\n    return 1.0 - x*x;\n}\n\nfunction NeuralNet() {\n\n    this.init = function(n1, n2, n3, n4) {\n\n        this.n1 = n1;\n        this.n2 = n2;\n        this.n3 = n3;\n        this.n4 = n4;\n\n        this.a1 = makeArray(this.n1, 1.0);\n        this.a2 = makeArray(this.n2, 1.0);\n        this.a3 = makeArray(this.n3, 1.0);\n        this.a4 = makeArray(this.n4, 1.0);\n\n        this.w12 = makeMatrix(this.n1, this.n2, 0.0);\n        this.w23 = makeMatrix(this.n2, this.n3, 0.0);\n        this.w34 = makeMatrix(this.n3, this.n4, 0.0);\n\n        for (var i = 0; i \u003c this.n1; i++)\n            for (var j = 0; j \u003c this.n2; j++)\n                this.w12[i][j] = rand(-0.3, 0.3);\n\n        for (var i = 0; i \u003c this.n2; i++)\n            for (var j = 0; j \u003c this.n3; j++)\n                this.w23[i][j] = rand(-0.3, 0.3);\n\n        for (var i = 0; i \u003c this.n3; i++)\n            for (var j = 0; j \u003c this.n4; j++)\n                this.w34[i][j] = rand(-0.3, 0.3);\n\n        this.c12 = makeMatrix(this.n1, this.n2, 0.0);\n        this.c23 = makeMatrix(this.n2, this.n3, 0.0);\n        this.c34 = makeMatrix(this.n3, this.n4, 0.0);\n\n        return this;\n    }\n\n    this.update = function(inputs) {\n        for (var i = 0; i \u003c this.n1; i++) {\n            this.a1[i] = inputs[i];\n        }\n        for (var j = 0; j \u003c this.n2; j++) {\n            var sum = 0;\n            for (var i = 0; i \u003c this.n1; i++) {\n                sum += this.a1[i]*this.w12[i][j];\n            }\n            this.a2[j] = sigmoid(sum);\n        }\n        for (var j = 0; j \u003c this.n3; j++) {\n            var sum = 0;\n            for (var i = 0; i \u003c this.n2; i++) {\n                sum += this.a2[i]*this.w23[i][j];\n            }\n            this.a3[j] = sigmoid(sum);\n        }\n        for (var j = 0; j \u003c this.n4; j++) {\n            var sum = 0;\n            for (var i = 0; i \u003c this.n3; i++) {\n                sum += this.a3[i]*this.w34[i][j];\n            }\n            \n            this.a4[j] = sigmoid(sum);\n        }\n\n        return this.a4;\n    }\n\n    this.backPropagate = function(targets, rate, moment) {\n\n        var deltas4 = makeArray(this.n4, 0);\n        for (var k = 0; k \u003c this.n4; k++) {\n            var error = targets[k]-this.a4[k];\n            deltas4[k] = dsigmoid(this.a4[k])*error;\n        }\n\n        var deltas3 = makeArray(this.n3, 0);\n        for (var j = 0; j \u003c this.n3; j++) {\n            var error = 0;\n            for (var k = 0; k \u003c this.n4; k++) {\n                error += deltas4[k]*this.w34[j][k];\n            }\n            deltas3[j] = dsigmoid(this.a3[j])*error;\n        }\n\n        var deltas2 = makeArray(this.n2, 0);\n        for (var j = 0; j \u003c this.n2; j++) {\n            var error = 0;\n            for (var k = 0; k \u003c this.n3; k++) {\n                error += deltas3[k]*this.w23[j][k];\n            }\n            deltas2[j] = dsigmoid(this.a2[j]) * error;\n        }\n\n\n        for (var j = 0; j \u003c this.n3; j++) {\n            for (var k = 0; k \u003c this.n4; k++) {\n                var change = deltas4[k]*this.a3[j];\n                this.w34[j][k] += rate*change + moment*this.c34[j][k];\n                this.c34[j][k] = change;\n            }\n        }\n        for (var j = 0; j \u003c this.n2; j++) {\n            for (var k = 0; k \u003c this.n3; k++) {\n                var change = deltas3[k]*this.a2[j];\n                this.w23[j][k] += rate*change + moment*this.c23[j][k];\n                this.c23[j][k] = change;\n            }\n        }\n        for (var j = 0; j \u003c this.n1; j++) {\n            for (var k = 0; k \u003c this.n2; k++) {\n                var change = deltas2[k]*this.a1[j];\n                this.w12[j][k] += rate*change + moment*this.c12[j][k];\n                this.c12[j][k] = change;\n            }\n        }\n    }\n\n\n    this.train = function(patterns, iterations, rate, moment) {\n        for (var i = 0; i \u003c iterations; i++) {\n            var error = 0.0;\n            for (var p in patterns) {\n                var pat = patterns[p];\n                var inputs = pat[0];\n                var targets = pat[1];\n                var outputs = this.update(inputs);\n                error = error + this.backPropagate(targets, rate, moment);\n            }\n        }\n    }\n}","description":null,"note":null,"status":"public","like_student_ids":[],"is_featured":false,"views":375,"hashid":"qmdsd5n3","is_content_changed":false,"review_status":"unsubmitted","submitted_at":null,"reviewed_at":null,"advise":null,"is_deleted":false}