readCSV(file, separator = “;”, header?)

This function reads a CSV file and returns a DataFrame object.

Parameters

• file: string: A string containing the file name of the CSV file to be read from.

• separator: string: A string representing the character which separates the CSV data. The default value is “;”.

• header: string[] | undefined: An optional array of column names for the DataFrame. If not provided, the first row of the CSV file is used as the header.

• Returns DataFrame: A DataFrame object.

• Exceptions Error: If the file name provided does not end with the .csv extension, an error is thrown.

Example

import * as ds from "https://deno.land/x/data_science@0.1.4/app.ts";

const df = ds.readCSV("data.csv", ",", ["col1", "col2", "col3"]);

This example reads a CSV file named data.csv, using , as the separator, and explicitly defines the column names as col1, col2, and col3. The resulting DataFrame object is stored in the df variable. If the file does not end with the .csv extension, an error is thrown. If the column names are not provided, the first row of the CSV file is used as the header.

Class: DataFrame

The DataFrame class represents a data frame object that can be used to manipulate and analyze tabular data.

Class Signature

export class DataFrame {
  public header: string[];

  constructor(name: string, data: string[][]);

  readCSV(file: string, seperator = ";", header?: string[]): DataFrame;
  getCol(name: string): Column;
  setCol(name: string, data: []): void;
  head(count: number): string[][];
}

Class Methods

constructor

Creates a new instance of the DataFrame class with the specified header and data.

constructor(header: string[], data: string[][]): DataFrame

Parameters

• header: string[] - Array of column names for the data frame
• data: string[][] - Array of data for the data frame

readCSV

Reads a csv file and returns a new instance of the DataFrame class with the specified header and data.

readCSV(file: string, seperator = ";", header?: string[]): DataFrame

Parameters

• file : string - File name of the csv file to read from
• seperator : string - Char, which seperates the csv data. Default value is ";"
• header : string[] - Array of column names for the data frame. Default value is the first row of the csv file

getCol

Returns the column with the specified name as a Column object.

getCol(name: string): Column

Parameters

• name : string - Name of the Column

setCol

Replaces a Column in the DataFrame with the specified data.

setCol(name: string, data: any[]): void

Parameters

• name : string - Name of the Column
• data : any[] - Data to replace in the DataFrame

head

Returns the specified number of rows from the top of the DataFrame as a two-dimensional array of strings.

head(count: number): string[][]

Parameters

• count : number - Number of rows to return

Example

import * as ds from "https://deno.land/x/data_science@0.1.4/app.ts";

// Read the CSV file and create a new DataFrame
const df = ds.DataFrame.readCSV("data.csv");

// Get the "age" column
const ageColumn = df.getCol("age");

// Print the first 5 values in the "age" column
console.log(ageColumn.data.slice(0, 5));

// Replace the "age" column with a new array of values
const newAgeColumn = [32, 43, 27, 19, 55];
df.setCol("age", newAgeColumn);

// Print the first 5 values in the "age" column again
console.log(df.getCol("age").head(5));

// Print the first 10 rows of the DataFrame
console.log(df.head(10));

Class: Column

This is a class that represents a column in a DataFrame. It stores the column name and an array of data.

Class Signature

export class Column {
  public name: string;
  public data: string[];

  constructor(name: string, data: string[]);

  apply(func: CallableFunction): void;
  head(count: number): string[];
}

Class Methods

constructor

This is the constructor method of the Column class.

constructor(name: string, data: string[]);

Parameters

• name : The name of the column.
• data : An array of data representing the column.

apply

This method applies a function to every item in the data array of the column.

apply(func: CallableFunction): void;

Parameters

• func : The function to be applied to each item in the data array.

head

Returns the amount of items given from the top

head(count: number): string[];

Parameters

• count : Number of items to return

Example

import * as ds from "https://deno.land/x/data_science@0.1.4/app.ts";

const df = ds.readCSV("fruitNames.csv");
const column = df.getCol("fruits");

console.log(column.head(5));
// Output: ["Apple", "Banana", "Cherry", "Date", "Elderberry"]

// Apply a function that capitalizes each item in the data array
column.apply((item: string) => item.toUpperCase());

console.log(column.head(5));
// Output: ["APPLE", "BANANA", "CHERRY", "DATE", "ELDERBERRY"]

Class: NeuralNetwork

The NeuralNetwork class is a JavaScript implementation of a basic feedforward neural network. It is capable of training on input-output data pairs and making predictions for new input data.

Constructor

Creates a new instance of the NeuralNetwork class.

constructor(layerSizes: number[] = [1, 1], outputLabels?: string[], activation: "SIGMOID" | "RELU" | "CAPPED RELU" | "TANH" = "SIGMOID", learningRate: number = 0.01)

Parameters

• layerSizes : an array of integers representing the number of nodes in each layer of the neural network. Must have at least two layers.
• outputLabels (optional): an array of strings representing the labels for the output layer. If not provided, the labels will be generated automatically based on the number of nodes in the output layer.
• activation (optional): a string representing the activation function to use for the neural network. Can be one of "SIGMOID", "RELU", "CAPPED RELU", or "TANH". Defaults to "SIGMOID".
• learningRate (optional): a number representing the learning rate to use for training the neural network. Defaults to 0.01.

Methods

train(x: number[], y: number[])

Trains the neural network on a single input-output data pair.

Parameters

• x: an array of numbers representing the input data.
• y: an array of numbers representing the expected output data.

fit(x: number[][], y: number[][])

Trains the neural network on multiple input-output data pairs.

Parameters

• x: an array of arrays of numbers representing the input data.
• y: an array of arrays of numbers representing the expected output data.

predict(input: number[])

Makes a prediction for a new input data.

Parameters

• input : an array of numbers representing the input data.

Returns

• A string representing the predicted output label.

Example

import * as ds from "https://deno.land/x/data_science@0.1.4/app.ts";

const nn = new ds.NeuralNetwork([2, 2, 1], ["0", "1"]);

const x = [[0,0],[0,1],[1,0],[1,1]];
const y = [[0,1],[0,1],[0,1],[0,1]];

nn.fit(x, y);

console.log(nn.predict([0, 0])); // Output: "0"
console.log(nn.predict([0, 1])); // Output: "1"
console.log(nn.predict([1, 0])); // Output: "1"
console.log(nn.predict([1, 1])); // Output: "0"