29:null 2a:["$","title",null,{"children":"cSP | Dominik's Workspace | Observable"}] 2f:T7c3,predData.length > 0 && Plot.plot({ marks: [ Plot.frame({ strokeOpacity: 0.1 }), ...(predTimelineSettings.show.includes("Signal") ? [Plot.lineY(predData[nPred - 1].signal)] : []), ...(predTimelineSettings.show.includes("Derivative") ? [ Plot.lineY(derivative(predData[nPred - 1].signal), { stroke: "blue" }) ] : []), Plot.rect( [ { onset: predData[nPred - 1].NN[0], offset: predData[nPred - 1].NN[1], y: [0, 1], type: "NN" }, { onset: predData[nPred - 1].NN_on.min, offset: predData[nPred - 1].NN_on.max, y: [0, 1], type: "NN uncertainty" }, { onset: predData[nPred - 1].NN_off.min, offset: predData[nPred - 1].NN_off.max, y: [0, 1], type: "NN uncertainty" }, { onset: manual[nPred - 1][0], offset: manual[nPred - 1][1], y: [-1, 0], type: "Manual" } ], { x1: (x) => x.onset * frequency, x2: (x) => x.offset * frequency, y1: (x) => x.y[0], y2: (x) => x.y[1], fill: "type", opacity: 0.2 } ) ], x: { type: "linear", domain: [0, predTimelineSettings.x], transform: (d) => d / frequency, label: "time (s)" }, y: { type: "linear", domain: [-0.005 / predTimelineSettings.y, 0.005 / predTimelineSettings.y] }, grid: true, color: { legend: true }, title: "Sample #" + predData[nPred - 1].id, subtitle: `cSP: NN=${(derivative(predData[nPred - 1].NN) * 1000).toFixed( 1 )}ms +- ${(predData[nPred - 1].NN_std * 1000).toFixed(1)}ms, Manual=${( derivative(manual[nPred - 1]) * 1000 ).toFixed(1)}ms`, width: 1000 })30:Tab8,randomTimeWarp = { function NaturalCubicSpline(knots, values) { const n = knots.length - 1, t = new Float64Array(n + 3), // knots y = new Float64Array(n + 3), // values s = new Float64Array(n + 2), // horizontal scales m = new Float64Array(n + 3), // slope at each knot d = new Float64Array(n + 2), // diagonal matrix coeffs = new Float64Array(4 * (n + 2)); // Chebyshev coefficients t.set(knots, 1), y.set(values, 1); // Natural cubic spline algorithm for (let i = 1; i < n + 1; i++) { s[i] = 1 / (t[i + 1] - t[i]); m[i] += m[i + 1] = 3 * s[i] * s[i] * (y[i + 1] - y[i]); } d[1] = 0.5 / s[1]; m[1] = d[1] * m[1]; for (let i = 2; i <= n + 1; i++) { d[i] = 1 / (2 * (s[i] + s[i - 1]) - s[i - 1] * s[i - 1] * d[i - 1]); m[i] = d[i] * (m[i] - s[i - 1] * m[i - 1]); } for (let i = n; i >= 1; i--) { m[i] -= d[i] * s[i] * m[i + 1]; } // Linear extrapolation t[0] = t[1] - 1; t[n + 2] = t[n + 1] + 1; y[0] = y[1] - m[1]; y[n + 2] = y[n + 1] + m[n + 1]; s[0] = s[n + 1] = 1; m[0] = m[1]; m[n + 2] = m[n + 1]; // Set up Chebyshev coefficients for (let i = 0; i < n + 2; i++) { const h = t[i + 1] - t[i]; const y0 = y[i], y1 = y[i + 1], m0 = h * m[i], m1 = h * m[i + 1], j = 4 * i; coeffs[j] = 0.5 * (y0 + y1 + 0.125 * (m0 - m1)); coeffs[j + 1] = 0.03125 * (18 * (y1 - y0) - m0 - m1); coeffs[j + 2] = 0.0625 * (m1 - m0); coeffs[j + 3] = 0.03125 * (2 * (y0 - y1) + m0 + m1); } return Object.assign( function eval_spline(u) { // Binary search let nn = t.length - 1, low = 0, half; while ((half = (nn / 2) | 0) > 0) { low += half * (t[low + half] <= u); nn -= half; } const i = low, j = 4 * i; u = (2 * u - t[i] - t[i + 1]) * s[i]; // scale to [–1, 1] // Clenshaw's method. let b1 = coeffs[j + 3], b2 = coeffs[j + 2] + 2 * u * b1; b1 = coeffs[j + 1] + 2 * u * b2 - b1; return coeffs[j] + u * b1 - b2; }, { knots: t, values: y, dvalues: m, scales: s, coeffs: coeffs } ); } return (args) => { const defaultArgs = { pts: 2, mag: 0.5, y0: 0, y1: 1 }; var { pts, mag, y0, y1 } = { ...defaultArgs, ...args }; const xs = d3.range(pts + 2).map((x) => x / (pts + 1)); // np.arange(pts+2, dtype='float') / (pts+1) mag = (mag * (y1 - y0)) / 2 / (pts + 1); const ys = xs.map( (x, i) => y0 + (y1 - y0) * x + (i > 0 && i < pts + 1 ? d3.randomUniform(-mag, mag)() : 0) ); return NaturalCubicSpline(xs, ys); }; }31:T7f7,scatter = (name) => { var labelmd = ""; var labelwt = ""; if (name == "NN") { const md = d3.rollup( cspTable.filter((x) => markers.includes(x["Onset/Offset/Duration"])), (v) => d3.mean(v, (x) => x[name + " Error"]), (x) => x["Split"] ); const wt = d3.rollup( cspTable.filter((x) => markers.includes(x["Onset/Offset/Duration"])), (v) => d3.mean(v, (x) => x[name + " Error"] <= scatterSettings.tol), (x) => x["Split"] ); labelmd = `${md.get("train").toFixed(1)}/${md.get("test").toFixed(1)}`; labelwt = `${(wt.get("train") * 100).toFixed(0)}/${( wt.get("test") * 100 ).toFixed(0)}`; } else { labelmd = d3.mean(cspTable, (x) => x[name + " Error"]).toFixed(1); labelwt = ( d3.mean(cspTable, (x) => x[name + " Error"] <= scatterSettings.tol) * 100 ).toFixed(0); } return Plot.plot({ x: { label: "Manual cSP", domain: [0, 280] }, y: { label: name + " cSP", domain: [0, 280] }, aspectRatio: 1, width: 400, grid: true, title: "cSP", symbol: { legend: true }, color: { legend: true }, subtitle: `mean deviation = ${labelmd}ms, within ${scatterSettings.tol}ms = ${labelwt}%`, marks: [ Plot.dot(cspTable, { y: (x) => x[name], x: (x) => x["Label"], sort: (x) => x.id == n, fill: (x) => (x.id == n ? "red" : "white"), stroke: name == "Julkunen" ? "black" : "Split", opacity: (x) => (x.id == n ? 1 : 0.3), filter: (x) => markers.includes(x["Onset/Offset/Duration"]), symbol: "Onset/Offset/Duration", r: 4 }), Plot.area([0.0, 300], { x1: (x) => x, y1: (x) => x - scatterSettings.tol, y2: (x) => x + scatterSettings.tol, opacity: 0.1 }), Plot.tip( cspTable, Plot.pointer({ y: (x) => x[name], x: (x) => x["Label"], title: (d) => `#${d.id} \n Manual: ${d["Label"]}ms \n ${name}: ${d[name]}ms` }) ) ] }); }32:T763,timeline = (name) => { const entry = Object.fromEntries( ["duration", "on", "off"].map((typ) => [ typ, cspTable.find((x) => x.id == n && x["Onset/Offset/Duration"] == typ) ]) ); return Plot.plot({ marks: [ Plot.frame({ strokeOpacity: 0.1 }), ...(timelineSettings.show.includes("Signal") ? [Plot.lineY(cspData[n - 1].signal)] : []), ...(timelineSettings.show.includes("Derivative") ? [Plot.lineY(derivative(cspData[n - 1].signal), { stroke: "blue" })] : []), Plot.ruleX([0]), ...(name == "Julkunen" && timelineSettings.show.includes("Derivative") ? [ Plot.ruleY([ (plot_julkunen_threshold / frequency) * 1000, (-plot_julkunen_threshold / frequency) * 1000 ]) ] : []), Plot.rect( [ { onset: entry.on["Label"], offset: entry.off["Label"], y: [0, 1], type: "Label" }, { onset: entry.on[name], offset: entry.off[name], y: [-1, 0], type: name == "NN" ? `${name} (${entry.on.Split})` : name } ], { x1: (x) => (x.onset * frequency) / 1000, x2: (x) => (x.offset * frequency) / 1000, y1: (x) => x.y[0], y2: (x) => x.y[1], fill: "type", opacity: 0.2 } ) ], x: { type: "linear", domain: [0, timelineSettings.x], transform: (d) => d / frequency, label: "time (s)" }, y: { type: "linear", domain: [-0.005 / timelineSettings.y, 0.005 / timelineSettings.y] }, grid: true, color: { legend: true }, title: "Sample #" + n, subtitle: `cSP: manual=${entry.duration["Label"]}ms, ${name}=${entry.duration[name]}ms` }); }33:[] 2d:{"id":"34cd7cc815b1602a","slug":null,"title":"cSP","thumbnail":"4c9829d31433c8d1f6c61b30f25086465226c75877469339ecc4008ed9945f2b","access_level":"public","create_time":"$D2023-09-14T07:42:58.371Z","update_time":"$D2023-10-12T14:44:22.306Z","view_time":null,"first_public_time":"$D2023-09-27T09:39:03.319Z","fork":false,"stars":0,"starred":false,"owner":{"type":"team","id":"6fbae63626342e49","login":"dominik2-ws","name":"Dominik's Workspace","avatar_url":"https://avatars-next.observableusercontent.com/29d54f7b0ada4485ab66a5b753687a16ebd840cee3486735bc72600f3928a351","tier":"starter_2024"},"version":898,"publish_version":null,"role":null,"authored":false,"authors":[{"type":"individual","id":"8d3a001f2b495452","login":"dominik2","name":"Dominik","avatar_url":"https://avatars-next.observableusercontent.com/2be6a595e5ee20eb59db0ea25f43e8e6d39f847748b68407f16331b31d5d6d82","tier":null}],"members":[{"type":"individual","id":"8d3a001f2b495452","login":"dominik2","name":"Dominik","avatar_url":"https://avatars-next.observableusercontent.com/2be6a595e5ee20eb59db0ea25f43e8e6d39f847748b68407f16331b31d5d6d82","tier":null,"role":"editor"},{"type":"team","id":"6fbae63626342e49","login":"dominik2-ws","name":"Dominik's Workspace","avatar_url":"https://avatars-next.observableusercontent.com/29d54f7b0ada4485ab66a5b753687a16ebd840cee3486735bc72600f3928a351","tier":"starter_2024","role":"editor"}],"comments":[{"id":"73c26ad526cddaff","reply_to_id":null,"content":"Just controls the width of the deviation plot in the scatter plots","create_time":"$D2023-09-14T21:11:40.574Z","update_time":"$D2023-09-14T21:12:03.720Z","resolved":false,"user":{"type":"individual","id":"8d3a001f2b495452","login":"dominik2","name":"Dominik","avatar_url":"https://avatars-next.observableusercontent.com/2be6a595e5ee20eb59db0ea25f43e8e6d39f847748b68407f16331b31d5d6d82","tier":null}}],"custom_thumbnail":false,"trashed":false,"subscribed":null,"forks":1,"collection":null,"type":"notebook","body":{"title":"cSP","cells":[{"id":0,"value":"# cSP","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":71,"value":"viewof n = Inputs.range([1, 200], {step: 1, label: \"Sample number\",value:1})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":327,"value":"viewof plot_julkunen_threshold = Inputs.range([.000001, .0001], {step: .000001, label: \"Julkunen threshold\",value: .00001})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":612,"value":"viewof timelineSettings = Inputs.form({\n x: Inputs.range([0.3, 2], { step: 0.05, label: \"x-Range (s)\", value: 0.5 }),\n y: Inputs.range([0.5, 100], { step: 0.1, label: \"y-Scale\", value: 1 }),\n show: Inputs.checkbox([\"Signal\", \"Derivative\"], {\n label: \"Plot\",\n value: [\"Signal\"]\n })\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":611,"value":"viewof scatterSettings = Inputs.form({\n tol: Inputs.range([1, 20], {\n step: 1,\n label: \"Duration tolerance (ms)\",\n value: 10\n }),\n show: Inputs.checkbox([\"Onset\", \"Offset\", \"Duration\"], {\n label: \"Plot\",\n value: [\"Duration\"]\n })\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":492,"value":"\n \n \n \n

\n ${timeline(\"NN\")}\n

\n ${scatter(\"NN\")}\n

\n \n","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined"},{"id":367,"value":"\n \n \n \n

\n ${timeline(\"Julkunen\")}\n

\n ${scatter(\"Julkunen\")}\n

\n \n","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined"},{"id":544,"value":"The NN algorithm performs very well on the training data, but to evaluate the performance one should only look at the testing data. The caption of the scatter plot shows the performance for training/testing separately.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":630,"value":"# Prediction","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":814,"value":"viewof pred_file = Inputs.file({ label: \"Unlabelled Signals\" })","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":677,"value":"viewof predTimelineSettings = Inputs.form({\n x: Inputs.range([0.2, 2], { step: 0.05, label: \"x-Range (s)\", value: 0.5 }),\n y: Inputs.range([0.5, 100], { step: 0.1, label: \"y-Scale\", value: 1 }),\n show: Inputs.checkbox([\"Signal\", \"Derivative\"], {\n label: \"Plot\",\n value: [\"Signal\"]\n })\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":662,"value":"viewof nPred = Inputs.range([1, predData.length], {\n step: 1,\n label: \"Prediction sample\",\n value: 1\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":713,"value":"predData.length > 0 &&\n bind(\n interval([0, 0.3], {\n label: \"Manual cSP\",\n step: 0.00001,\n format: ([start, end]) =>\n `${(start * 1000).toFixed(1)} – ${(end * 1000).toFixed(1)}ms (NN: ${(\n predData[nPred - 1].NN[0] * 1000\n ).toFixed(1)} – ${(predData[nPred - 1].NN[1] * 1000).toFixed(1)}ms)`,\n width: 850\n }),\n viewof manual,\n nPred - 1\n )","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":655,"value":"$2f","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":853,"value":"Inputs.table(predLabels, {\n columns: [\"id\", \"NN\", \"NN_std\", \"Manual\", \"Correction\"],\n header: {\n id: \"Sample\",\n NN_std: \"NN Uncertainty\"\n },\n format: {\n NN: (x) =>\n `${((x[1] - x[0]) * 1000).toFixed(1)}ms (${(x[0] * 1000).toFixed(1)}-${(\n x[1] * 1000\n ).toFixed(0)}ms)`,\n Manual: (x) =>\n `${((x[1] - x[0]) * 1000).toFixed(1)}ms (${(x[0] * 1000).toFixed(1)}-${(\n x[1] * 1000\n ).toFixed(0)}ms)`,\n NN_std: (x) => (x * 1000).toFixed(1),\n Correction: (x) => (x * 1000).toFixed(1)\n }\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":877,"value":"DOM.download(\n new Blob([JSON.stringify(predLabels, null, 2)], {\n type: \"application/json\"\n }),\n \"predictions.json\",\n \"Download predictions.json\"\n)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":132,"value":"---\n# Appendix","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":607,"value":"cspData = {\n const request = await fetch(\"https://wirhabenzeit.github.io/csp/data.json\");\n return request.json();\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":606,"value":"Inputs.table(cspData, {\n columns: [\"id\", \"split\", \"csp_label\"],\n header: {\n id: \"Sample\",\n split: \"Type\",\n csp_label: \"CSP (Label)\"\n },\n format: {\n csp_label: (x) =>\n `${((x[1] - x[0]) * 1000).toFixed()}ms (${(x[0] * 1000).toFixed(0)}-${(\n x[1] * 1000\n ).toFixed(0)}ms)`\n }\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":68,"value":"frequency = 3000","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":605,"value":"## Julkunen Prediction","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":259,"value":"derivative = (signal) =>\n signal.slice(1, signal.length).map((x, i) => x - signal[i])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":263,"value":"julkunenPrediction = (\n threshold = 0.0225,\n freq = 3000,\n minLen = 0.03,\n maxTime = 0.3\n ) =>\n (signal) => {\n const diff = derivative(signal).map((x) => Math.abs(x * freq * 1e-3));\n const minInd = Math.round(minLen * freq);\n const windowsOfSmallDerivative = [];\n var [i, j] = [0, 0];\n while (j < Math.min(diff.length, maxTime * freq)) {\n if (Math.abs(diff[j]) >= threshold) {\n if (j - i > minInd) windowsOfSmallDerivative.push([i, j - 1]);\n i = j + 1;\n j = i;\n } else j += 1;\n }\n if (j - i > minInd) windowsOfSmallDerivative.push([i, j - 1]);\n if (windowsOfSmallDerivative.length == 0) return [0, 0];\n return windowsOfSmallDerivative\n .sort((a, b) => d3.descending(a[1] - a[0], b[1] - b[0]))[0]\n .map((x) => x / freq);\n }","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":395,"value":"labelsJulkunen = new d3.InternMap(\n cspData.map((x) => [\n x.id,\n julkunenPrediction(plot_julkunen_threshold)(x.signal)\n ])\n)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":602,"value":"## Neural Network Prediction","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":601,"value":"modelDimension = 256","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":617,"value":"modelPrediction = {\n const ort =\n await require(\"https://cdn.jsdelivr.net/npm/onnxruntime-web@1.15.1/dist/ort.js\");\n ort.env.wasm.wasmPaths =\n \"https://cdn.jsdelivr.net/npm/onnxruntime-web@1.15.1/dist/\";\n const executionProviders = [\"wasm\"];\n const session = await ort.InferenceSession.create(\n \"https://wirhabenzeit.github.io/csp/xresnet.onnx\",\n { executionProviders }\n );\n return async (signals) => {\n const bs = signals.length;\n const length = signals[0].length;\n const results = await session.run({\n input: new ort.Tensor(\"float32\", signals.flat(), [bs, 1, length])\n });\n const arr = Array.from(results.output.data);\n return d3.range(bs).map((i) => arr.slice(2 * i, 2 * i + 2));\n };\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":599,"value":"$30","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":598,"value":"randomWarpedSignals = (signal, copies, twArgs) => {\n return d3.range(copies).map(() => {\n const tw = randomTimeWarp(twArgs);\n const mwHelper = randomTimeWarp({\n ...(\"pts\" in twArgs ? { pts: twArgs.pts } : {}),\n ...(\"mag\" in twArgs ? { mag: twArgs.mag } : {})\n });\n const mw = (t) => 1 + mwHelper(t) - t;\n const sig = Array.from(\n d3\n .range(modelDimension)\n .map(\n (i) =>\n 1000 *\n signal[Math.round(tw(i / (modelDimension - 1)) * frequency)] *\n mw(i / (modelDimension - 1))\n )\n );\n return { tw, sig };\n });\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":597,"value":"augmentedModelPrediction = async (signal, copies, twArgs) => {\n const augSignals = randomWarpedSignals(signal, copies, twArgs);\n const rawResults = await modelPrediction(augSignals.map((x) => x.sig));\n const transformedResults = rawResults.map((vec, i) =>\n vec.map(augSignals[i].tw)\n );\n const funcs = {\n on: (x) => x[0],\n off: (x) => x[1],\n duration: (x) => x[1] - x[0]\n };\n const stats = {\n mean: d3.mean,\n median: d3.median,\n std: d3.deviation,\n min: d3.min,\n max: d3.max\n };\n return Object.fromEntries(\n Object.entries(funcs).map(([label, func]) => [\n label,\n Object.fromEntries(\n Object.entries(stats).map(([stat, statFunc]) => [\n stat,\n statFunc(transformedResults, func)\n ])\n )\n ])\n );\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":596,"value":"labelsNN = {\n const labels = await Promise.all(\n cspData.map((x) =>\n augmentedModelPrediction(x.signal, 5, {\n pts: 3,\n mag: 0.2,\n y0: 0,\n y1: 0.4\n })\n )\n );\n return new d3.InternMap(cspData.map((x, i) => [x.id, labels[i]]));\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":595,"value":"## Data with predicitons ","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":594,"value":"cspLabels = {\n const arr = [];\n const funcs = {\n on: (x) => x[0],\n off: (x) => x[1],\n duration: (x) => x[1] - x[0]\n };\n cspData.forEach((x) =>\n Object.entries(funcs).forEach(([lab, func]) =>\n arr.push({\n id: x.id,\n split: x.split,\n event: lab,\n label: Math.round(func(x.csp_label) * 10000) / 10,\n julkunen: Math.round(func(labelsJulkunen.get(x.id)) * 10000) / 10,\n meanNN: Math.round(labelsNN.get(x.id)[lab].median * 10000) / 10,\n stdNN: Math.round(labelsNN.get(x.id)[lab].std * 10000) / 10\n })\n )\n );\n return arr;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":593,"value":"cspTable = cspLabels","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":592,"value":"## Plots","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":708,"value":"import { interval } from \"@mootari/range-slider\"","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":831,"value":"function disposal(element) {\n return new Promise((resolve) => {\n requestAnimationFrame(() => {\n const target = element.closest(\".observablehq\");\n if (!target) return resolve();\n const observer = new MutationObserver(() => {\n if (target.contains(element)) return;\n observer.disconnect(), resolve();\n });\n observer.observe(target, { childList: true });\n });\n });\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":829,"value":"function bind(target, source, arg, invalidation = disposal(target)) {\n const sourceEvent = \"input\";\n const onsource = () => {\n target.value = source.value[arg];\n };\n const ontarget = () => {\n source.value[arg] = target.value;\n source.dispatchEvent(new Event(sourceEvent, { bubbles: true }));\n };\n onsource();\n target.addEventListener(\"input\", ontarget);\n source.addEventListener(sourceEvent, onsource);\n invalidation.then(() => source.removeEventListener(sourceEvent, onsource));\n return target;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":750,"value":"function set(input, value) {\n input.value = value;\n input.dispatchEvent(new Event(\"input\", { bubbles: true }));\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":591,"value":"markers = [\n ...(scatterSettings.show.includes(\"Duration\") ? [\"duration\"] : []),\n ...(scatterSettings.show.includes(\"Onset\") ? [\"on\"] : []),\n ...(scatterSettings.show.includes(\"Offset\") ? [\"off\"] : [])\n]","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":590,"value":"$31","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":589,"value":"$32","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":882,"value":"## Prediction","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":637,"value":"predData = {\n if (pred_file == null) return [];\n const data = await pred_file.json();\n const NN_pred = await Promise.all(\n data.map((d) =>\n augmentedModelPrediction(d.signal, 7, {\n y1: 0.3,\n mag: 0.2,\n pts: 3\n })\n )\n );\n data.forEach((d, i) => {\n d.NN = [NN_pred[i].on.mean, NN_pred[i].off.mean];\n d.NN_std = NN_pred[i].duration.std;\n d.NN_on = NN_pred[i].on;\n d.NN_off = NN_pred[i].off;\n });\n data.sort((x, y) => -x.NN_std + y.NN_std);\n return data;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":835,"value":"viewof manual = Inputs.input(predData.map((d) => d.NN))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":857,"value":"predLabels = predData.map((d, i) => ({\n id: d.id,\n NN: d.NN,\n NN_std: d.NN_std,\n Manual: manual[i],\n Correction: Math.abs(d.NN[1] - d.NN[0] - manual[i][1] + manual[i][0])\n}))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":895,"value":"examplePlot = timeline(\"NN\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"}],"files":"$W33","stdlib":"1","titling":"auto"}} 2e:"$2d:body"