feat:node-modules

node_modules/mathjs/lib/cjs/function/statistics/corr.js

@@ -0,0 +1,90 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createCorr = void 0;
+var _factory = require("../../utils/factory.js");
+const name = 'corr';
+const dependencies = ['typed', 'matrix', 'mean', 'sqrt', 'sum', 'add', 'subtract', 'multiply', 'pow', 'divide'];
+const createCorr = exports.createCorr = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    matrix,
+    sqrt,
+    sum,
+    add,
+    subtract,
+    multiply,
+    pow,
+    divide
+  } = _ref;
+  /**
+   * Compute the correlation coefficient of a two list with values, For matrices, the matrix correlation coefficient is calculated.
+   *
+   * Syntax:
+   *
+   *     math.corr(A, B)
+   *
+   * Examples:
+   *
+   *     math.corr([1, 2, 3, 4, 5], [4, 5, 6, 7, 8])     // returns 1
+   *     math.corr([1, 2.2, 3, 4.8, 5], [4, 5.3, 6.6, 7, 8])     //returns 0.9569941688503644
+   *     math.corr([[1, 2.2, 3, 4.8, 5], [4, 5.3, 6.6, 7, 8]],[[1, 2.2, 3, 4.8, 5], [4, 5.3, 6.6, 7, 8]])   // returns [1,1]
+   *
+   * See also:
+   *
+   *     median, mean, min, max, sum, prod, std, variance
+   *
+   * @param {Array | Matrix} A The first array or matrix to compute correlation coefficient
+   * @param {Array | Matrix} B The second array or matrix to compute correlation coefficient
+   * @return {*} The correlation coefficient
+   */
+  return typed(name, {
+    'Array, Array': function (A, B) {
+      return _corr(A, B);
+    },
+    'Matrix, Matrix': function (A, B) {
+      const res = _corr(A.toArray(), B.toArray());
+      return Array.isArray(res) ? matrix(res) : res;
+    }
+  });
+  /**
+   * Calculate the correlation coefficient between two arrays or matrices.
+   * @param {Array | Matrix} A
+   * @param {Array | Matrix} B
+   * @return {*} correlation coefficient
+   * @private
+   */
+  function _corr(A, B) {
+    const correlations = [];
+    if (Array.isArray(A[0]) && Array.isArray(B[0])) {
+      if (A.length !== B.length) {
+        throw new SyntaxError('Dimension mismatch. Array A and B must have the same length.');
+      }
+      for (let i = 0; i < A.length; i++) {
+        if (A[i].length !== B[i].length) {
+          throw new SyntaxError('Dimension mismatch. Array A and B must have the same number of elements.');
+        }
+        correlations.push(correlation(A[i], B[i]));
+      }
+      return correlations;
+    } else {
+      if (A.length !== B.length) {
+        throw new SyntaxError('Dimension mismatch. Array A and B must have the same number of elements.');
+      }
+      return correlation(A, B);
+    }
+  }
+  function correlation(A, B) {
+    const n = A.length;
+    const sumX = sum(A);
+    const sumY = sum(B);
+    const sumXY = A.reduce((acc, x, index) => add(acc, multiply(x, B[index])), 0);
+    const sumXSquare = sum(A.map(x => pow(x, 2)));
+    const sumYSquare = sum(B.map(y => pow(y, 2)));
+    const numerator = subtract(multiply(n, sumXY), multiply(sumX, sumY));
+    const denominator = sqrt(multiply(subtract(multiply(n, sumXSquare), pow(sumX, 2)), subtract(multiply(n, sumYSquare), pow(sumY, 2))));
+    return divide(numerator, denominator);
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/cumsum.js

@@ -0,0 +1,129 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createCumSum = void 0;
+var _collection = require("../../utils/collection.js");
+var _factory = require("../../utils/factory.js");
+var _switch2 = require("../../utils/switch.js");
+var _improveErrorMessage = require("./utils/improveErrorMessage.js");
+var _array = require("../../utils/array.js");
+var _IndexError = require("../../error/IndexError.js");
+const name = 'cumsum';
+const dependencies = ['typed', 'add', 'unaryPlus'];
+const createCumSum = exports.createCumSum = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    add,
+    unaryPlus
+  } = _ref;
+  /**
+   * Compute the cumulative sum of a matrix or a list with values.
+   * In case of a (multi dimensional) array or matrix, the cumulative sums
+   * along a specified dimension (defaulting to the first) will be calculated.
+   *
+   * Syntax:
+   *
+   *     math.cumsum(a, b, c, ...)
+   *     math.cumsum(A)
+   *
+   * Examples:
+   *
+   *     math.cumsum(2, 1, 4, 3)               // returns [2, 3, 7, 10]
+   *     math.cumsum([2, 1, 4, 3])             // returns [2, 3, 7, 10]
+   *     math.cumsum([[1, 2], [3, 4]])         // returns [[1, 2], [4, 6]]
+   *     math.cumsum([[1, 2], [3, 4]], 0)      // returns [[1, 2], [4, 6]]
+   *     math.cumsum([[1, 2], [3, 4]], 1)      // returns [[1, 3], [3, 7]]
+   *     math.cumsum([[2, 5], [4, 3], [1, 7]]) // returns [[2, 5], [6, 8], [7, 15]]
+   *
+   * See also:
+   *
+   *    mean, median, min, max, prod, std, variance, sum
+   *
+   * @param {... *} args  A single matrix or or multiple scalar values
+   * @return {*} The cumulative sum of all values
+   */
+  return typed(name, {
+    // sum([a, b, c, d, ...])
+    Array: _cumsum,
+    Matrix: function (matrix) {
+      return matrix.create(_cumsum(matrix.valueOf(), matrix.datatype()));
+    },
+    // sum([a, b, c, d, ...], dim)
+    'Array, number | BigNumber': _ncumSumDim,
+    'Matrix, number | BigNumber': function (matrix, dim) {
+      return matrix.create(_ncumSumDim(matrix.valueOf(), dim), matrix.datatype());
+    },
+    // cumsum(a, b, c, d, ...)
+    '...': function (args) {
+      if ((0, _collection.containsCollections)(args)) {
+        throw new TypeError('All values expected to be scalar in function cumsum');
+      }
+      return _cumsum(args);
+    }
+  });
+
+  /**
+     * Recursively calculate the cumulative sum of an n-dimensional array
+     * @param {Array} array
+     * @return {number} cumsum
+     * @private
+     */
+  function _cumsum(array) {
+    try {
+      return _cumsummap(array);
+    } catch (err) {
+      throw (0, _improveErrorMessage.improveErrorMessage)(err, name);
+    }
+  }
+  function _cumsummap(array) {
+    if (array.length === 0) {
+      return [];
+    }
+    const sums = [unaryPlus(array[0])]; // unaryPlus converts to number if need be
+    for (let i = 1; i < array.length; ++i) {
+      // Must use add below and not addScalar for the case of summing a
+      // 2+-dimensional array along the 0th dimension (the row vectors,
+      // or higher-d analogues, are literally added to each other).
+      sums.push(add(sums[i - 1], array[i]));
+    }
+    return sums;
+  }
+  function _ncumSumDim(array, dim) {
+    const size = (0, _array.arraySize)(array);
+    if (dim < 0 || dim >= size.length) {
+      // TODO: would be more clear when throwing a DimensionError here
+      throw new _IndexError.IndexError(dim, size.length);
+    }
+    try {
+      return _cumsumDimensional(array, dim);
+    } catch (err) {
+      throw (0, _improveErrorMessage.improveErrorMessage)(err, name);
+    }
+  }
+
+  /* Possible TODO: Refactor _reduce in collection.js to be able to work here as well */
+  function _cumsumDimensional(mat, dim) {
+    let i, ret, tran;
+    if (dim <= 0) {
+      const initialValue = mat[0][0];
+      if (!Array.isArray(initialValue)) {
+        return _cumsummap(mat);
+      } else {
+        tran = (0, _switch2._switch)(mat);
+        ret = [];
+        for (i = 0; i < tran.length; i++) {
+          ret[i] = _cumsumDimensional(tran[i], dim - 1);
+        }
+        return ret;
+      }
+    } else {
+      ret = [];
+      for (i = 0; i < mat.length; i++) {
+        ret[i] = _cumsumDimensional(mat[i], dim - 1);
+      }
+      return ret;
+    }
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/mad.js

@@ -0,0 +1,70 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createMad = void 0;
+var _array = require("../../utils/array.js");
+var _factory = require("../../utils/factory.js");
+var _improveErrorMessage = require("./utils/improveErrorMessage.js");
+const name = 'mad';
+const dependencies = ['typed', 'abs', 'map', 'median', 'subtract'];
+const createMad = exports.createMad = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    abs,
+    map,
+    median,
+    subtract
+  } = _ref;
+  /**
+   * Compute the median absolute deviation of a matrix or a list with values.
+   * The median absolute deviation is defined as the median of the absolute
+   * deviations from the median.
+   *
+   * Syntax:
+   *
+   *     math.mad(a, b, c, ...)
+   *     math.mad(A)
+   *
+   * Examples:
+   *
+   *     math.mad(10, 20, 30)             // returns 10
+   *     math.mad([1, 2, 3])              // returns 1
+   *     math.mad([[1, 2, 3], [4, 5, 6]]) // returns 1.5
+   *
+   * See also:
+   *
+   *     median, mean, std, abs
+   *
+   * @param {Array | Matrix} array
+   *                        A single matrix or multiple scalar values.
+   * @return {*} The median absolute deviation.
+   */
+  return typed(name, {
+    // mad([a, b, c, d, ...])
+    'Array | Matrix': _mad,
+    // mad(a, b, c, d, ...)
+    '...': function (args) {
+      return _mad(args);
+    }
+  });
+  function _mad(array) {
+    array = (0, _array.flatten)(array.valueOf());
+    if (array.length === 0) {
+      throw new Error('Cannot calculate median absolute deviation (mad) of an empty array');
+    }
+    try {
+      const med = median(array);
+      return median(map(array, function (value) {
+        return abs(subtract(value, med));
+      }));
+    } catch (err) {
+      if (err instanceof TypeError && err.message.includes('median')) {
+        throw new TypeError(err.message.replace('median', 'mad'));
+      } else {
+        throw (0, _improveErrorMessage.improveErrorMessage)(err, 'mad');
+      }
+    }
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/max.js

@@ -0,0 +1,111 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createMax = void 0;
+var _collection = require("../../utils/collection.js");
+var _factory = require("../../utils/factory.js");
+var _number = require("../../utils/number.js");
+var _improveErrorMessage = require("./utils/improveErrorMessage.js");
+const name = 'max';
+const dependencies = ['typed', 'config', 'numeric', 'larger'];
+const createMax = exports.createMax = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    config,
+    numeric,
+    larger
+  } = _ref;
+  /**
+   * Compute the maximum value of a matrix or a  list with values.
+   * In case of a multidimensional array, the maximum of the flattened array
+   * will be calculated. When `dim` is provided, the maximum over the selected
+   * dimension will be calculated. Parameter `dim` is zero-based.
+   *
+   * Syntax:
+   *
+   *     math.max(a, b, c, ...)
+   *     math.max(A)
+   *     math.max(A, dimension)
+   *
+   * Examples:
+   *
+   *     math.max(2, 1, 4, 3)                  // returns 4
+   *     math.max([2, 1, 4, 3])                // returns 4
+   *
+   *     // maximum over a specified dimension (zero-based)
+   *     math.max([[2, 5], [4, 3], [1, 7]], 0) // returns [4, 7]
+   *     math.max([[2, 5], [4, 3], [1, 7]], 1) // returns [5, 4, 7]
+   *
+   *     math.max(2.7, 7.1, -4.5, 2.0, 4.1)    // returns 7.1
+   *     math.min(2.7, 7.1, -4.5, 2.0, 4.1)    // returns -4.5
+   *
+   * See also:
+   *
+   *    mean, median, min, prod, std, sum, variance
+   *
+   * @param {... *} args  A single matrix or or multiple scalar values
+   * @return {*} The maximum value
+   */
+  return typed(name, {
+    // max([a, b, c, d, ...])
+    'Array | Matrix': _max,
+    // max([a, b, c, d, ...], dim)
+    'Array | Matrix, number | BigNumber': function (array, dim) {
+      return (0, _collection.reduce)(array, dim.valueOf(), _largest);
+    },
+    // max(a, b, c, d, ...)
+    '...': function (args) {
+      if ((0, _collection.containsCollections)(args)) {
+        throw new TypeError('Scalar values expected in function max');
+      }
+      return _max(args);
+    }
+  });
+
+  /**
+   * Return the largest of two values
+   * @param {*} x
+   * @param {*} y
+   * @returns {*} Returns x when x is largest, or y when y is largest
+   * @private
+   */
+  function _largest(x, y) {
+    try {
+      return larger(x, y) ? x : y;
+    } catch (err) {
+      throw (0, _improveErrorMessage.improveErrorMessage)(err, 'max', y);
+    }
+  }
+
+  /**
+   * Recursively calculate the maximum value in an n-dimensional array
+   * @param {Array} array
+   * @return {number} max
+   * @private
+   */
+  function _max(array) {
+    let res;
+    (0, _collection.deepForEach)(array, function (value) {
+      try {
+        if (isNaN(value) && typeof value === 'number') {
+          res = NaN;
+        } else if (res === undefined || larger(value, res)) {
+          res = value;
+        }
+      } catch (err) {
+        throw (0, _improveErrorMessage.improveErrorMessage)(err, 'max', value);
+      }
+    });
+    if (res === undefined) {
+      throw new Error('Cannot calculate max of an empty array');
+    }
+
+    // make sure returning numeric value: parse a string into a numeric value
+    if (typeof res === 'string') {
+      res = numeric(res, (0, _number.safeNumberType)(res, config));
+    }
+    return res;
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/mean.js

@@ -0,0 +1,100 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createMean = void 0;
+var _collection = require("../../utils/collection.js");
+var _array = require("../../utils/array.js");
+var _factory = require("../../utils/factory.js");
+var _improveErrorMessage = require("./utils/improveErrorMessage.js");
+const name = 'mean';
+const dependencies = ['typed', 'add', 'divide'];
+const createMean = exports.createMean = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    add,
+    divide
+  } = _ref;
+  /**
+   * Compute the mean value of matrix or a list with values.
+   * In case of a multidimensional array, the mean of the flattened array
+   * will be calculated. When `dim` is provided, the maximum over the selected
+   * dimension will be calculated. Parameter `dim` is zero-based.
+   *
+   * Syntax:
+   *
+   *     math.mean(a, b, c, ...)
+   *     math.mean(A)
+   *     math.mean(A, dimension)
+   *
+   * Examples:
+   *
+   *     math.mean(2, 1, 4, 3)                     // returns 2.5
+   *     math.mean([1, 2.7, 3.2, 4])               // returns 2.725
+   *
+   *     math.mean([[2, 5], [6, 3], [1, 7]], 0)    // returns [3, 5]
+   *     math.mean([[2, 5], [6, 3], [1, 7]], 1)    // returns [3.5, 4.5, 4]
+   *
+   * See also:
+   *
+   *     median, min, max, sum, prod, std, variance
+   *
+   * @param {... *} args  A single matrix or or multiple scalar values
+   * @return {*} The mean of all values
+   */
+  return typed(name, {
+    // mean([a, b, c, d, ...])
+    'Array | Matrix': _mean,
+    // mean([a, b, c, d, ...], dim)
+    'Array | Matrix, number | BigNumber': _nmeanDim,
+    // mean(a, b, c, d, ...)
+    '...': function (args) {
+      if ((0, _collection.containsCollections)(args)) {
+        throw new TypeError('Scalar values expected in function mean');
+      }
+      return _mean(args);
+    }
+  });
+
+  /**
+   * Calculate the mean value in an n-dimensional array, returning a
+   * n-1 dimensional array
+   * @param {Array} array
+   * @param {number} dim
+   * @return {number} mean
+   * @private
+   */
+  function _nmeanDim(array, dim) {
+    try {
+      const sum = (0, _collection.reduce)(array, dim, add);
+      const s = Array.isArray(array) ? (0, _array.arraySize)(array) : array.size();
+      return divide(sum, s[dim]);
+    } catch (err) {
+      throw (0, _improveErrorMessage.improveErrorMessage)(err, 'mean');
+    }
+  }
+
+  /**
+   * Recursively calculate the mean value in an n-dimensional array
+   * @param {Array} array
+   * @return {number} mean
+   * @private
+   */
+  function _mean(array) {
+    let sum;
+    let num = 0;
+    (0, _collection.deepForEach)(array, function (value) {
+      try {
+        sum = sum === undefined ? value : add(sum, value);
+        num++;
+      } catch (err) {
+        throw (0, _improveErrorMessage.improveErrorMessage)(err, 'mean', value);
+      }
+    });
+    if (num === 0) {
+      throw new Error('Cannot calculate the mean of an empty array');
+    }
+    return divide(sum, num);
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/median.js

@@ -0,0 +1,114 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createMedian = void 0;
+var _collection = require("../../utils/collection.js");
+var _array = require("../../utils/array.js");
+var _factory = require("../../utils/factory.js");
+var _improveErrorMessage = require("./utils/improveErrorMessage.js");
+const name = 'median';
+const dependencies = ['typed', 'add', 'divide', 'compare', 'partitionSelect'];
+const createMedian = exports.createMedian = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    add,
+    divide,
+    compare,
+    partitionSelect
+  } = _ref;
+  /**
+   * Recursively calculate the median of an n-dimensional array
+   * @param {Array} array
+   * @return {Number} median
+   * @private
+   */
+  function _median(array) {
+    try {
+      array = (0, _array.flatten)(array.valueOf());
+      const num = array.length;
+      if (num === 0) {
+        throw new Error('Cannot calculate median of an empty array');
+      }
+      if (num % 2 === 0) {
+        // even: return the average of the two middle values
+        const mid = num / 2 - 1;
+        const right = partitionSelect(array, mid + 1);
+
+        // array now partitioned at mid + 1, take max of left part
+        let left = array[mid];
+        for (let i = 0; i < mid; ++i) {
+          if (compare(array[i], left) > 0) {
+            left = array[i];
+          }
+        }
+        return middle2(left, right);
+      } else {
+        // odd: return the middle value
+        const m = partitionSelect(array, (num - 1) / 2);
+        return middle(m);
+      }
+    } catch (err) {
+      throw (0, _improveErrorMessage.improveErrorMessage)(err, 'median');
+    }
+  }
+
+  // helper function to type check the middle value of the array
+  const middle = typed({
+    'number | BigNumber | Complex | Unit': function (value) {
+      return value;
+    }
+  });
+
+  // helper function to type check the two middle value of the array
+  const middle2 = typed({
+    'number | BigNumber | Complex | Unit, number | BigNumber | Complex | Unit': function (left, right) {
+      return divide(add(left, right), 2);
+    }
+  });
+
+  /**
+   * Compute the median of a matrix or a list with values. The values are
+   * sorted and the middle value is returned. In case of an even number of
+   * values, the average of the two middle values is returned.
+   * Supported types of values are: Number, BigNumber, Unit
+   *
+   * In case of a (multi dimensional) array or matrix, the median of all
+   * elements will be calculated.
+   *
+   * Syntax:
+   *
+   *     math.median(a, b, c, ...)
+   *     math.median(A)
+   *
+   * Examples:
+   *
+   *     math.median(5, 2, 7)        // returns 5
+   *     math.median([3, -1, 5, 7])  // returns 4
+   *
+   * See also:
+   *
+   *     mean, min, max, sum, prod, std, variance, quantileSeq
+   *
+   * @param {... *} args  A single matrix or or multiple scalar values
+   * @return {*} The median
+   */
+  return typed(name, {
+    // median([a, b, c, d, ...])
+    'Array | Matrix': _median,
+    // median([a, b, c, d, ...], dim)
+    'Array | Matrix, number | BigNumber': function (array, dim) {
+      // TODO: implement median(A, dim)
+      throw new Error('median(A, dim) is not yet supported');
+      // return reduce(arguments[0], arguments[1], ...)
+    },
+    // median(a, b, c, d, ...)
+    '...': function (args) {
+      if ((0, _collection.containsCollections)(args)) {
+        throw new TypeError('Scalar values expected in function median');
+      }
+      return _median(args);
+    }
+  });
+});

node_modules/mathjs/lib/cjs/function/statistics/min.js

@@ -0,0 +1,111 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createMin = void 0;
+var _collection = require("../../utils/collection.js");
+var _factory = require("../../utils/factory.js");
+var _number = require("../../utils/number.js");
+var _improveErrorMessage = require("./utils/improveErrorMessage.js");
+const name = 'min';
+const dependencies = ['typed', 'config', 'numeric', 'smaller'];
+const createMin = exports.createMin = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    config,
+    numeric,
+    smaller
+  } = _ref;
+  /**
+   * Compute the minimum value of a matrix or a  list of values.
+   * In case of a multidimensional array, the minimum of the flattened array
+   * will be calculated. When `dim` is provided, the minimum over the selected
+   * dimension will be calculated. Parameter `dim` is zero-based.
+   *
+   * Syntax:
+   *
+   *     math.min(a, b, c, ...)
+   *     math.min(A)
+   *     math.min(A, dimension)
+   *
+   * Examples:
+   *
+   *     math.min(2, 1, 4, 3)                  // returns 1
+   *     math.min([2, 1, 4, 3])                // returns 1
+   *
+   *     // minimum over a specified dimension (zero-based)
+   *     math.min([[2, 5], [4, 3], [1, 7]], 0) // returns [1, 3]
+   *     math.min([[2, 5], [4, 3], [1, 7]], 1) // returns [2, 3, 1]
+   *
+   *     math.max(2.7, 7.1, -4.5, 2.0, 4.1)    // returns 7.1
+   *     math.min(2.7, 7.1, -4.5, 2.0, 4.1)    // returns -4.5
+   *
+   * See also:
+   *
+   *    mean, median, max, prod, std, sum, variance
+   *
+   * @param {... *} args  A single matrix or or multiple scalar values
+   * @return {*} The minimum value
+   */
+  return typed(name, {
+    // min([a, b, c, d, ...])
+    'Array | Matrix': _min,
+    // min([a, b, c, d, ...], dim)
+    'Array | Matrix, number | BigNumber': function (array, dim) {
+      return (0, _collection.reduce)(array, dim.valueOf(), _smallest);
+    },
+    // min(a, b, c, d, ...)
+    '...': function (args) {
+      if ((0, _collection.containsCollections)(args)) {
+        throw new TypeError('Scalar values expected in function min');
+      }
+      return _min(args);
+    }
+  });
+
+  /**
+   * Return the smallest of two values
+   * @param {*} x
+   * @param {*} y
+   * @returns {*} Returns x when x is smallest, or y when y is smallest
+   * @private
+   */
+  function _smallest(x, y) {
+    try {
+      return smaller(x, y) ? x : y;
+    } catch (err) {
+      throw (0, _improveErrorMessage.improveErrorMessage)(err, 'min', y);
+    }
+  }
+
+  /**
+   * Recursively calculate the minimum value in an n-dimensional array
+   * @param {Array} array
+   * @return {number} min
+   * @private
+   */
+  function _min(array) {
+    let min;
+    (0, _collection.deepForEach)(array, function (value) {
+      try {
+        if (isNaN(value) && typeof value === 'number') {
+          min = NaN;
+        } else if (min === undefined || smaller(value, min)) {
+          min = value;
+        }
+      } catch (err) {
+        throw (0, _improveErrorMessage.improveErrorMessage)(err, 'min', value);
+      }
+    });
+    if (min === undefined) {
+      throw new Error('Cannot calculate min of an empty array');
+    }
+
+    // make sure returning numeric value: parse a string into a numeric value
+    if (typeof min === 'string') {
+      min = numeric(min, (0, _number.safeNumberType)(min, config));
+    }
+    return min;
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/mode.js

@@ -0,0 +1,82 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createMode = void 0;
+var _array = require("../../utils/array.js");
+var _factory = require("../../utils/factory.js");
+const name = 'mode';
+const dependencies = ['typed', 'isNaN', 'isNumeric'];
+const createMode = exports.createMode = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    isNaN,
+    isNumeric
+  } = _ref;
+  /**
+  * Computes the mode of a set of numbers or a list with values(numbers or characters).
+  * If there are multiple modes, it returns a list of those values.
+  *
+  * Syntax:
+  *
+  *     math.mode(a, b, c, ...)
+  *     math.mode(A)
+  *
+  * Examples:
+  *
+  *     math.mode(2, 1, 4, 3, 1)                            // returns [1]
+  *     math.mode([1, 2.7, 3.2, 4, 2.7])                    // returns [2.7]
+  *     math.mode(1, 4, 6, 1, 6)                             // returns [1, 6]
+  *     math.mode('a','a','b','c')                           // returns ["a"]
+  *     math.mode(1, 1.5, 'abc')                             // returns [1, 1.5, "abc"]
+  *
+  * See also:
+  *
+  *     median,
+  *     mean
+  *
+  * @param {... *} args  A single matrix
+  * @return {*} The mode of all values
+  */
+  return typed(name, {
+    'Array | Matrix': _mode,
+    '...': function (args) {
+      return _mode(args);
+    }
+  });
+
+  /**
+   * Calculates the mode in an 1-dimensional array
+   * @param {Array} values
+   * @return {Array} mode
+   * @private
+   */
+  function _mode(values) {
+    values = (0, _array.flatten)(values.valueOf());
+    const num = values.length;
+    if (num === 0) {
+      throw new Error('Cannot calculate mode of an empty array');
+    }
+    const count = {};
+    let mode = [];
+    let max = 0;
+    for (let i = 0; i < values.length; i++) {
+      const value = values[i];
+      if (isNumeric(value) && isNaN(value)) {
+        throw new Error('Cannot calculate mode of an array containing NaN values');
+      }
+      if (!(value in count)) {
+        count[value] = 0;
+      }
+      count[value]++;
+      if (count[value] === max) {
+        mode.push(value);
+      } else if (count[value] > max) {
+        max = count[value];
+        mode = [value];
+      }
+    }
+    return mode;
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/prod.js

@@ -0,0 +1,85 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createProd = void 0;
+var _collection = require("../../utils/collection.js");
+var _factory = require("../../utils/factory.js");
+var _number = require("../../utils/number.js");
+var _improveErrorMessage = require("./utils/improveErrorMessage.js");
+const name = 'prod';
+const dependencies = ['typed', 'config', 'multiplyScalar', 'numeric'];
+const createProd = exports.createProd = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    config,
+    multiplyScalar,
+    numeric
+  } = _ref;
+  /**
+   * Compute the product of a matrix or a list with values.
+   * In case of a multidimensional array or matrix, the sum of all
+   * elements will be calculated.
+   *
+   * Syntax:
+   *
+   *     math.prod(a, b, c, ...)
+   *     math.prod(A)
+   *
+   * Examples:
+   *
+   *     math.multiply(2, 3)           // returns 6
+   *     math.prod(2, 3)               // returns 6
+   *     math.prod(2, 3, 4)            // returns 24
+   *     math.prod([2, 3, 4])          // returns 24
+   *     math.prod([[2, 5], [4, 3]])   // returns 120
+   *
+   * See also:
+   *
+   *    mean, median, min, max, sum, std, variance
+   *
+   * @param {... *} args  A single matrix or or multiple scalar values
+   * @return {*} The product of all values
+   */
+  return typed(name, {
+    // prod([a, b, c, d, ...])
+    'Array | Matrix': _prod,
+    // prod([a, b, c, d, ...], dim)
+    'Array | Matrix, number | BigNumber': function (array, dim) {
+      // TODO: implement prod(A, dim)
+      throw new Error('prod(A, dim) is not yet supported');
+      // return reduce(arguments[0], arguments[1], math.prod)
+    },
+    // prod(a, b, c, d, ...)
+    '...': function (args) {
+      return _prod(args);
+    }
+  });
+
+  /**
+   * Recursively calculate the product of an n-dimensional array
+   * @param {Array} array
+   * @return {number} prod
+   * @private
+   */
+  function _prod(array) {
+    let prod;
+    (0, _collection.deepForEach)(array, function (value) {
+      try {
+        prod = prod === undefined ? value : multiplyScalar(prod, value);
+      } catch (err) {
+        throw (0, _improveErrorMessage.improveErrorMessage)(err, 'prod', value);
+      }
+    });
+
+    // make sure returning numeric value: parse a string into a numeric value
+    if (typeof prod === 'string') {
+      prod = numeric(prod, (0, _number.safeNumberType)(prod, config));
+    }
+    if (prod === undefined) {
+      throw new Error('Cannot calculate prod of an empty array');
+    }
+    return prod;
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/quantileSeq.js

@@ -0,0 +1,172 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createQuantileSeq = void 0;
+var _is = require("../../utils/is.js");
+var _array = require("../../utils/array.js");
+var _factory = require("../../utils/factory.js");
+var _apply = require("../matrix/apply.js");
+const name = 'quantileSeq';
+const dependencies = ['typed', '?bignumber', 'add', 'subtract', 'divide', 'multiply', 'partitionSelect', 'compare', 'isInteger', 'smaller', 'smallerEq', 'larger'];
+const createQuantileSeq = exports.createQuantileSeq = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    bignumber,
+    add,
+    subtract,
+    divide,
+    multiply,
+    partitionSelect,
+    compare,
+    isInteger,
+    smaller,
+    smallerEq,
+    larger
+  } = _ref;
+  const apply = (0, _apply.createApply)({
+    typed,
+    isInteger
+  });
+
+  /**
+   * Compute the prob order quantile of a matrix or a list with values.
+   * The sequence is sorted and the middle value is returned.
+   * Supported types of sequence values are: Number, BigNumber, Unit
+   * Supported types of probability are: Number, BigNumber
+   *
+   * In case of a multidimensional array or matrix, the prob order quantile
+   * of all elements will be calculated.
+   *
+   * Syntax:
+   *
+   *     math.quantileSeq(A, prob[, sorted])
+   *     math.quantileSeq(A, [prob1, prob2, ...][, sorted])
+   *     math.quantileSeq(A, N[, sorted])
+   *
+   * Examples:
+   *
+   *     math.quantileSeq([3, -1, 5, 7], 0.5)         // returns 4
+   *     math.quantileSeq([3, -1, 5, 7], [1/3, 2/3])  // returns [3, 5]
+   *     math.quantileSeq([3, -1, 5, 7], 2)           // returns [3, 5]
+   *     math.quantileSeq([-1, 3, 5, 7], 0.5, true)   // returns 4
+   *
+   * See also:
+   *
+   *     median, mean, min, max, sum, prod, std, variance
+   *
+   * @param {Array, Matrix} data                A single matrix or Array
+   * @param {Number, BigNumber, Array} probOrN  prob is the order of the quantile, while N is
+   *                                            the amount of evenly distributed steps of
+   *                                            probabilities; only one of these options can
+   *                                            be provided
+   * @param {Boolean} sorted=false              is data sorted in ascending order
+   * @return {Number, BigNumber, Unit, Array}   Quantile(s)
+   */
+  return typed(name, {
+    'Array | Matrix, number | BigNumber': (data, p) => _quantileSeqProbNumber(data, p, false),
+    'Array | Matrix, number | BigNumber, number': (data, prob, dim) => _quantileSeqDim(data, prob, false, dim, _quantileSeqProbNumber),
+    'Array | Matrix, number | BigNumber, boolean': _quantileSeqProbNumber,
+    'Array | Matrix, number | BigNumber, boolean, number': (data, prob, sorted, dim) => _quantileSeqDim(data, prob, sorted, dim, _quantileSeqProbNumber),
+    'Array | Matrix, Array | Matrix': (data, p) => _quantileSeqProbCollection(data, p, false),
+    'Array | Matrix, Array | Matrix, number': (data, prob, dim) => _quantileSeqDim(data, prob, false, dim, _quantileSeqProbCollection),
+    'Array | Matrix, Array | Matrix, boolean': _quantileSeqProbCollection,
+    'Array | Matrix, Array | Matrix, boolean, number': (data, prob, sorted, dim) => _quantileSeqDim(data, prob, sorted, dim, _quantileSeqProbCollection)
+  });
+  function _quantileSeqDim(data, prob, sorted, dim, fn) {
+    return apply(data, dim, x => fn(x, prob, sorted));
+  }
+  function _quantileSeqProbNumber(data, probOrN, sorted) {
+    let probArr;
+    const dataArr = data.valueOf();
+    if (smaller(probOrN, 0)) {
+      throw new Error('N/prob must be non-negative');
+    }
+    if (smallerEq(probOrN, 1)) {
+      // quantileSeq([a, b, c, d, ...], prob[,sorted])
+      return (0, _is.isNumber)(probOrN) ? _quantileSeq(dataArr, probOrN, sorted) : bignumber(_quantileSeq(dataArr, probOrN, sorted));
+    }
+    if (larger(probOrN, 1)) {
+      // quantileSeq([a, b, c, d, ...], N[,sorted])
+      if (!isInteger(probOrN)) {
+        throw new Error('N must be a positive integer');
+      }
+
+      // largest possible Array length is 2^32-1
+      // 2^32 < 10^15, thus safe conversion guaranteed
+      if (larger(probOrN, 4294967295)) {
+        throw new Error('N must be less than or equal to 2^32-1, as that is the maximum length of an Array');
+      }
+      const nPlusOne = add(probOrN, 1);
+      probArr = [];
+      for (let i = 0; smaller(i, probOrN); i++) {
+        const prob = divide(i + 1, nPlusOne);
+        probArr.push(_quantileSeq(dataArr, prob, sorted));
+      }
+      return (0, _is.isNumber)(probOrN) ? probArr : bignumber(probArr);
+    }
+  }
+
+  /**
+   * Calculate the prob order quantile of an n-dimensional array.
+   *
+   * @param {Array, Matrix} array
+   * @param {Array, Matrix} prob
+   * @param {Boolean} sorted
+   * @return {Number, BigNumber, Unit} prob order quantile
+   * @private
+   */
+
+  function _quantileSeqProbCollection(data, probOrN, sorted) {
+    const dataArr = data.valueOf();
+    // quantileSeq([a, b, c, d, ...], [prob1, prob2, ...][,sorted])
+    const probOrNArr = probOrN.valueOf();
+    const probArr = [];
+    for (let i = 0; i < probOrNArr.length; ++i) {
+      probArr.push(_quantileSeq(dataArr, probOrNArr[i], sorted));
+    }
+    return probArr;
+  }
+
+  /**
+   * Calculate the prob order quantile of an n-dimensional array.
+   *
+   * @param {Array} array
+   * @param {Number, BigNumber} prob
+   * @param {Boolean} sorted
+   * @return {Number, BigNumber, Unit} prob order quantile
+   * @private
+   */
+  function _quantileSeq(array, prob, sorted) {
+    const flat = (0, _array.flatten)(array);
+    const len = flat.length;
+    if (len === 0) {
+      throw new Error('Cannot calculate quantile of an empty sequence');
+    }
+    const index = (0, _is.isNumber)(prob) ? prob * (len - 1) : prob.times(len - 1);
+    const integerPart = (0, _is.isNumber)(prob) ? Math.floor(index) : index.floor().toNumber();
+    const fracPart = (0, _is.isNumber)(prob) ? index % 1 : index.minus(integerPart);
+    if (isInteger(index)) {
+      return sorted ? flat[index] : partitionSelect(flat, (0, _is.isNumber)(prob) ? index : index.valueOf());
+    }
+    let left;
+    let right;
+    if (sorted) {
+      left = flat[integerPart];
+      right = flat[integerPart + 1];
+    } else {
+      right = partitionSelect(flat, integerPart + 1);
+
+      // max of partition is kth largest
+      left = flat[integerPart];
+      for (let i = 0; i < integerPart; ++i) {
+        if (compare(flat[i], left) > 0) {
+          left = flat[i];
+        }
+      }
+    }
+    // Q(prob) = (1-f)*A[floor(index)] + f*A[floor(index)+1]
+    return add(multiply(left, subtract(1, fracPart)), multiply(right, fracPart));
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/std.js

@@ -0,0 +1,103 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createStd = void 0;
+var _factory = require("../../utils/factory.js");
+var _is = require("../../utils/is.js");
+const name = 'std';
+const dependencies = ['typed', 'map', 'sqrt', 'variance'];
+const createStd = exports.createStd = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    map,
+    sqrt,
+    variance
+  } = _ref;
+  /**
+   * Compute the standard deviation of a matrix or a  list with values.
+   * The standard deviations is defined as the square root of the variance:
+   * `std(A) = sqrt(variance(A))`.
+   * In case of a (multi dimensional) array or matrix, the standard deviation
+   * over all elements will be calculated by default, unless an axis is specified
+   * in which case the standard deviation will be computed along that axis.
+   *
+   * Additionally, it is possible to compute the standard deviation along the rows
+   * or columns of a matrix by specifying the dimension as the second argument.
+   *
+   * Optionally, the type of normalization can be specified as the final
+   * parameter. The parameter `normalization` can be one of the following values:
+   *
+   * - 'unbiased' (default) The sum of squared errors is divided by (n - 1)
+   * - 'uncorrected'        The sum of squared errors is divided by n
+   * - 'biased'             The sum of squared errors is divided by (n + 1)
+   *
+   *
+   * Syntax:
+   *
+   *     math.std(a, b, c, ...)
+   *     math.std(A)
+   *     math.std(A, normalization)
+   *     math.std(A, dimension)
+   *     math.std(A, dimension, normalization)
+   *
+   * Examples:
+   *
+   *     math.std(2, 4, 6)                     // returns 2
+   *     math.std([2, 4, 6, 8])                // returns 2.581988897471611
+   *     math.std([2, 4, 6, 8], 'uncorrected') // returns 2.23606797749979
+   *     math.std([2, 4, 6, 8], 'biased')      // returns 2
+   *
+   *     math.std([[1, 2, 3], [4, 5, 6]])      // returns 1.8708286933869707
+   *     math.std([[1, 2, 3], [4, 6, 8]], 0)    // returns [2.1213203435596424, 2.8284271247461903, 3.5355339059327378]
+   *     math.std([[1, 2, 3], [4, 6, 8]], 1)    // returns [1, 2]
+   *     math.std([[1, 2, 3], [4, 6, 8]], 1, 'biased') // returns [0.7071067811865476, 1.4142135623730951]
+   *
+   * See also:
+   *
+   *    mean, median, max, min, prod, sum, variance
+   *
+   * @param {Array | Matrix} array
+   *                        A single matrix or or multiple scalar values
+   * @param {string} [normalization='unbiased']
+   *                        Determines how to normalize the variance.
+   *                        Choose 'unbiased' (default), 'uncorrected', or 'biased'.
+   * @param dimension {number | BigNumber}
+   *                        Determines the axis to compute the standard deviation for a matrix
+   * @return {*} The standard deviation
+   */
+  return typed(name, {
+    // std([a, b, c, d, ...])
+    'Array | Matrix': _std,
+    // std([a, b, c, d, ...], normalization)
+    'Array | Matrix, string': _std,
+    // std([a, b, c, c, ...], dim)
+    'Array | Matrix, number | BigNumber': _std,
+    // std([a, b, c, c, ...], dim, normalization)
+    'Array | Matrix, number | BigNumber, string': _std,
+    // std(a, b, c, d, ...)
+    '...': function (args) {
+      return _std(args);
+    }
+  });
+  function _std(array, normalization) {
+    if (array.length === 0) {
+      throw new SyntaxError('Function std requires one or more parameters (0 provided)');
+    }
+    try {
+      const v = variance.apply(null, arguments);
+      if ((0, _is.isCollection)(v)) {
+        return map(v, sqrt);
+      } else {
+        return sqrt(v);
+      }
+    } catch (err) {
+      if (err instanceof TypeError && err.message.includes(' variance')) {
+        throw new TypeError(err.message.replace(' variance', ' std'));
+      } else {
+        throw err;
+      }
+    }
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/sum.js

@@ -0,0 +1,91 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createSum = void 0;
+var _collection = require("../../utils/collection.js");
+var _factory = require("../../utils/factory.js");
+var _number = require("../../utils/number.js");
+var _improveErrorMessage = require("./utils/improveErrorMessage.js");
+const name = 'sum';
+const dependencies = ['typed', 'config', 'add', 'numeric'];
+const createSum = exports.createSum = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    config,
+    add,
+    numeric
+  } = _ref;
+  /**
+   * Compute the sum of a matrix or a list with values.
+   * In case of a multidimensional array or matrix, the sum of all
+   * elements will be calculated.
+   *
+   * Syntax:
+   *
+   *     math.sum(a, b, c, ...)
+   *     math.sum(A)
+   *     math.sum(A, dimension)
+   *
+   * Examples:
+   *
+   *     math.sum(2, 1, 4, 3)               // returns 10
+   *     math.sum([2, 1, 4, 3])             // returns 10
+   *     math.sum([[2, 5], [4, 3], [1, 7]]) // returns 22
+   *
+   * See also:
+   *
+   *    mean, median, min, max, prod, std, variance, cumsum
+   *
+   * @param {... *} args  A single matrix or multiple scalar values
+   * @return {*} The sum of all values
+   */
+  return typed(name, {
+    // sum([a, b, c, d, ...])
+    'Array | Matrix': _sum,
+    // sum([a, b, c, d, ...], dim)
+    'Array | Matrix, number | BigNumber': _nsumDim,
+    // sum(a, b, c, d, ...)
+    '...': function (args) {
+      if ((0, _collection.containsCollections)(args)) {
+        throw new TypeError('Scalar values expected in function sum');
+      }
+      return _sum(args);
+    }
+  });
+
+  /**
+   * Recursively calculate the sum of an n-dimensional array
+   * @param {Array | Matrix} array
+   * @return {number} sum
+   * @private
+   */
+  function _sum(array) {
+    let sum;
+    (0, _collection.deepForEach)(array, function (value) {
+      try {
+        sum = sum === undefined ? value : add(sum, value);
+      } catch (err) {
+        throw (0, _improveErrorMessage.improveErrorMessage)(err, 'sum', value);
+      }
+    });
+
+    // make sure returning numeric value: parse a string into a numeric value
+    if (sum === undefined) {
+      sum = numeric(0, config.number);
+    }
+    if (typeof sum === 'string') {
+      sum = numeric(sum, (0, _number.safeNumberType)(sum, config));
+    }
+    return sum;
+  }
+  function _nsumDim(array, dim) {
+    try {
+      const sum = (0, _collection.reduce)(array, dim, add);
+      return sum;
+    } catch (err) {
+      throw (0, _improveErrorMessage.improveErrorMessage)(err, 'sum');
+    }
+  }
+});

node_modules/mathjs/lib/cjs/function/statistics/utils/improveErrorMessage.js

@@ -0,0 +1,30 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.improveErrorMessage = improveErrorMessage;
+var _is = require("../../../utils/is.js");
+/**
+ * Improve error messages for statistics functions. Errors are typically
+ * thrown in an internally used function like larger, causing the error
+ * not to mention the function (like max) which is actually used by the user.
+ *
+ * @param {Error} err
+ * @param {String} fnName
+ * @param {*} [value]
+ * @return {Error}
+ */
+function improveErrorMessage(err, fnName, value) {
+  // TODO: add information with the index (also needs transform in expression parser)
+  let details;
+  if (String(err).includes('Unexpected type')) {
+    details = arguments.length > 2 ? ' (type: ' + (0, _is.typeOf)(value) + ', value: ' + JSON.stringify(value) + ')' : ' (type: ' + err.data.actual + ')';
+    return new TypeError('Cannot calculate ' + fnName + ', unexpected type of argument' + details);
+  }
+  if (String(err).includes('complex numbers')) {
+    details = arguments.length > 2 ? ' (type: ' + (0, _is.typeOf)(value) + ', value: ' + JSON.stringify(value) + ')' : '';
+    return new TypeError('Cannot calculate ' + fnName + ', no ordering relation is defined for complex numbers' + details);
+  }
+  return err;
+}

node_modules/mathjs/lib/cjs/function/statistics/variance.js

@@ -0,0 +1,159 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.createVariance = void 0;
+var _collection = require("../../utils/collection.js");
+var _is = require("../../utils/is.js");
+var _factory = require("../../utils/factory.js");
+var _improveErrorMessage = require("./utils/improveErrorMessage.js");
+const DEFAULT_NORMALIZATION = 'unbiased';
+const name = 'variance';
+const dependencies = ['typed', 'add', 'subtract', 'multiply', 'divide', 'apply', 'isNaN'];
+const createVariance = exports.createVariance = /* #__PURE__ */(0, _factory.factory)(name, dependencies, _ref => {
+  let {
+    typed,
+    add,
+    subtract,
+    multiply,
+    divide,
+    apply,
+    isNaN
+  } = _ref;
+  /**
+   * Compute the variance of a matrix or a  list with values.
+   * In case of a multidimensional array or matrix, the variance over all
+   * elements will be calculated.
+   *
+   * Additionally, it is possible to compute the variance along the rows
+   * or columns of a matrix by specifying the dimension as the second argument.
+   *
+   * Optionally, the type of normalization can be specified as the final
+   * parameter. The parameter `normalization` can be one of the following values:
+   *
+   * - 'unbiased' (default) The sum of squared errors is divided by (n - 1)
+   * - 'uncorrected'        The sum of squared errors is divided by n
+   * - 'biased'             The sum of squared errors is divided by (n + 1)
+   *
+   *
+   * Note that older browser may not like the variable name `var`. In that
+   * case, the function can be called as `math['var'](...)` instead of
+   * `math.var(...)`.
+   *
+   * Syntax:
+   *
+   *     math.variance(a, b, c, ...)
+   *     math.variance(A)
+   *     math.variance(A, normalization)
+   *     math.variance(A, dimension)
+   *     math.variance(A, dimension, normalization)
+   *
+   * Examples:
+   *
+   *     math.variance(2, 4, 6)                     // returns 4
+   *     math.variance([2, 4, 6, 8])                // returns 6.666666666666667
+   *     math.variance([2, 4, 6, 8], 'uncorrected') // returns 5
+   *     math.variance([2, 4, 6, 8], 'biased')      // returns 4
+   *
+   *     math.variance([[1, 2, 3], [4, 5, 6]])      // returns 3.5
+   *     math.variance([[1, 2, 3], [4, 6, 8]], 0)   // returns [4.5, 8, 12.5]
+   *     math.variance([[1, 2, 3], [4, 6, 8]], 1)   // returns [1, 4]
+   *     math.variance([[1, 2, 3], [4, 6, 8]], 1, 'biased') // returns [0.5, 2]
+   *
+   * See also:
+   *
+   *    mean, median, max, min, prod, std, sum
+   *
+   * @param {Array | Matrix} array
+   *                        A single matrix or or multiple scalar values
+   * @param {string} [normalization='unbiased']
+   *                        Determines how to normalize the variance.
+   *                        Choose 'unbiased' (default), 'uncorrected', or 'biased'.
+   * @param dimension {number | BigNumber}
+   *                        Determines the axis to compute the variance for a matrix
+   * @return {*} The variance
+   */
+  return typed(name, {
+    // variance([a, b, c, d, ...])
+    'Array | Matrix': function (array) {
+      return _var(array, DEFAULT_NORMALIZATION);
+    },
+    // variance([a, b, c, d, ...], normalization)
+    'Array | Matrix, string': _var,
+    // variance([a, b, c, c, ...], dim)
+    'Array | Matrix, number | BigNumber': function (array, dim) {
+      return _varDim(array, dim, DEFAULT_NORMALIZATION);
+    },
+    // variance([a, b, c, c, ...], dim, normalization)
+    'Array | Matrix, number | BigNumber, string': _varDim,
+    // variance(a, b, c, d, ...)
+    '...': function (args) {
+      return _var(args, DEFAULT_NORMALIZATION);
+    }
+  });
+
+  /**
+   * Recursively calculate the variance of an n-dimensional array
+   * @param {Array} array
+   * @param {string} normalization
+   *                        Determines how to normalize the variance:
+   *                        - 'unbiased'    The sum of squared errors is divided by (n - 1)
+   *                        - 'uncorrected' The sum of squared errors is divided by n
+   *                        - 'biased'      The sum of squared errors is divided by (n + 1)
+   * @return {number | BigNumber} variance
+   * @private
+   */
+  function _var(array, normalization) {
+    let sum;
+    let num = 0;
+    if (array.length === 0) {
+      throw new SyntaxError('Function variance requires one or more parameters (0 provided)');
+    }
+
+    // calculate the mean and number of elements
+    (0, _collection.deepForEach)(array, function (value) {
+      try {
+        sum = sum === undefined ? value : add(sum, value);
+        num++;
+      } catch (err) {
+        throw (0, _improveErrorMessage.improveErrorMessage)(err, 'variance', value);
+      }
+    });
+    if (num === 0) throw new Error('Cannot calculate variance of an empty array');
+    const mean = divide(sum, num);
+
+    // calculate the variance
+    sum = undefined;
+    (0, _collection.deepForEach)(array, function (value) {
+      const diff = subtract(value, mean);
+      sum = sum === undefined ? multiply(diff, diff) : add(sum, multiply(diff, diff));
+    });
+    if (isNaN(sum)) {
+      return sum;
+    }
+    switch (normalization) {
+      case 'uncorrected':
+        return divide(sum, num);
+      case 'biased':
+        return divide(sum, num + 1);
+      case 'unbiased':
+        {
+          const zero = (0, _is.isBigNumber)(sum) ? sum.mul(0) : 0;
+          return num === 1 ? zero : divide(sum, num - 1);
+        }
+      default:
+        throw new Error('Unknown normalization "' + normalization + '". ' + 'Choose "unbiased" (default), "uncorrected", or "biased".');
+    }
+  }
+  function _varDim(array, dim, normalization) {
+    try {
+      if (array.length === 0) {
+        throw new SyntaxError('Function variance requires one or more parameters (0 provided)');
+      }
+      return apply(array, dim, x => _var(x, normalization));
+    } catch (err) {
+      throw (0, _improveErrorMessage.improveErrorMessage)(err, 'variance');
+    }
+  }
+});