mathias-brandewinder

#r "office.dll"
#r "Microsoft.Office.Interop.Excel"
open Microsoft.Office.Interop.Excel
open System.Runtime.InteropServices
let xl = Marshal.GetActiveObject("Excel.Application") :?> Microsoft.Office.Interop.Excel.Application
let wbs = xl.Workbooks
let wb = wbs.[1]
let sh = wb.Worksheets.[1] :?> Worksheet
sh.Cells.[1,1] <- "Hello from F#!"

mathias-brandewinder

type City = { Name: string; Lat: float; Lng: float }

let cities =
    [ ("San Francisco", "CA", "United States");
      ("New York", "NY", "United States");
      ("Hoboken", "NJ", "United States");
      ("Redmond", "WA", "United States");
      ("Seattle", "WA", "United States");
      ("Boston", "MA", "United States");
      ("Boston", "MA", "United States");

mathias-brandewinder

#I @"C:\Program Files (x86)\Reference Assemblies\Microsoft\FSharp\3.0\Runtime\v4.0\Type Providers"
open System
#r "System.ServiceModel.dll"
open Microsoft.FSharp.Linq
#r "FSharp.Data.TypeProviders"

type terraService = Microsoft.FSharp.Data.TypeProviders.WsdlService<"http://terraserver-usa.com/TerraService2.asmx?WSDL">

let terraClient = terraService.GetTerraServiceSoap ()

mathias-brandewinder

// This F# dojo is directly inspired by the 
// Digit Recognizer competition from Kaggle.com:
// http://www.kaggle.com/c/digit-recognizer
// The datasets below are simply shorter versions of
// the training dataset from Kaggle.
 
// The goal of the dojo will be to
// create a classifier that uses training data
// to recognize hand-written digits, and
// evaluate the quality of our classifier

mathias-brandewinder

open System
open System.IO

let path = @"C:\Users\Mathias\Desktop\Dojo\DigitsSample.csv"
let data = File.ReadAllLines(path)
let parsed = data |> Array.map (fun line -> line.Split(','))
let parsed2 = parsed.[1..] |> Array.map (fun line -> line |> Array.map (fun x -> Convert.ToInt32(x)))

type Example = { Number:int; Pixels:int[] }
let sample = parsed2 |> Array.map (fun x -> { Number = x.[0]; Pixels = x.[1..] })

mathias-brandewinder

namespace Discretization

// Recursive minimal entropy partitioning,
// based on Fayyad & Irani 1993. 
// See the following article, section 3.3,
// for a description of the algorithm:
// http://www.math.unipd.it/~dulli/corso04/disc.pdf
// Note: this can certainly be optimized.
module MDL =

mathias-brandewinder

open System

// Sample with replacement
let replaceSampler (rng: Random) (counts: int[]) =
    let N = counts |> Array.sum
    let draw = rng.Next(N)
    let rec find index cumul =
        let n = counts.[index]
        if draw < (cumul + n) 
        then index 

mathias-brandewinder

namespace FSharpTests

module Analytics = 

    open System

    let getData (skuId: int) =
        // fake data for now:
        // a list of (date, value) observations
        [   DateTime(2010, 1, 1), 10.;

mathias-brandewinder

This dojo is directly inspired by the Digit Recognizer competition from Kaggle.com:
http://www.kaggle.com/c/digit-recognizer
The datasets below are simply shorter versions of the training dataset from Kaggle.


The dataset
*************

2 datasets can be downloaded here:

mathias-brandewinder

#r @"..\packages\Accord.2.8.1.0\lib\Accord.dll"
#r @"..\packages\Accord.Math.2.8.1.0\lib\Accord.Math.dll"
#r @"..\packages\Accord.Statistics.2.8.1.0\lib\Accord.Statistics.dll"
#r @"..\packages\Accord.MachineLearning.2.8.1.0\lib\Accord.MachineLearning.dll"

open System
open System.IO

open Accord.MachineLearning
open Accord.MachineLearning.VectorMachines