ML.NET SERIES: AutoML untuk Ranking

Hi rekan coder, 2020 sebentar lagi selesai, semoga pandemi segera usai. 

Sambil mengisi waktu dengan hal-hal positif mari kita sama-sama belajar menggunakan fitur baru ML.NET yaitu AutoML for ranking. Nah biasanya kalau kita mau kasih rekomendasi ke orang, kita lihat orang-orang pada ngasih rating yang bagus ga ke barang-barang tertentu. Nah secara sederhana itu yang secara nalar mudah dimengerti. 

Nah sekarang algorima ranking bisa dilakukan training secara otomatis dengan AutoML, artinya rekan-rekan tinggal siapin datasetnya saja, nanti biar ML.NET yang carikan model paling akurat dalam durasi waktu tertentu. 

Sebagai contoh kita punya informasi mengenai search index, rating, dan usernya. Berikut sample codenya:

using System;
using System.Collections.Generic;
using System.Text;
using System.IO;
using System.Linq;
using Microsoft.ML.Data;
using Microsoft.ML;
using Microsoft.ML.AutoML;

namespace DetectWeatherAnomaly
{
public class AutoRanking
{
private static string ModelPath = "Model.zip";
private static string LabelColumnName = "Label";
private static string GroupColumnName = "GroupId";
private static uint ExperimentTime = 60;

public static void Run()
{
MLContext mlContext = new MLContext();

// STEP 1: Load data
IDataView trainDataView = mlContext.Data.LoadFromEnumerable<SearchData>(GenerateData());
IDataView testDataView = mlContext.Data.LoadFromEnumerable <SearchData>(GenerateData(10));

// STEP 2: Run AutoML experiment
Console.WriteLine($"Running AutoML recommendation experiment for {ExperimentTime} seconds...");
ExperimentResult<RankingMetrics> experimentResult = mlContext.Auto()
.CreateRankingExperiment(new RankingExperimentSettings() { MaxExperimentTimeInSeconds = ExperimentTime })
.Execute(trainDataView, testDataView,
new ColumnInformation()
{
LabelColumnName = LabelColumnName,
GroupIdColumnName = GroupColumnName
});

// STEP 3: Print metric from best model
RunDetail<RankingMetrics> bestRun = experimentResult.BestRun;
Console.WriteLine($"Total models produced: {experimentResult.RunDetails.Count()}");
Console.WriteLine($"Best model's trainer: {bestRun.TrainerName}");
Console.WriteLine($"Metrics of best model from validation data --");
PrintMetrics(bestRun.ValidationMetrics);

// STEP 5: Evaluate test data
IDataView testDataViewWithBestScore = bestRun.Model.Transform(testDataView);
RankingMetrics testMetrics = mlContext.Ranking.Evaluate(testDataViewWithBestScore, labelColumnName: LabelColumnName);
Console.WriteLine($"Metrics of best model on test data --");
PrintMetrics(testMetrics);

// STEP 6: Save the best model for later deployment and inferencing
mlContext.Model.Save(bestRun.Model, trainDataView.Schema, ModelPath);

// STEP 7: Create prediction engine from the best trained model
var predictionEngine = mlContext.Model.CreatePredictionEngine<SearchData, SearchDataPrediction>(bestRun.Model);

// STEP 8: Initialize a new test, and get the prediction
var testPage = new SearchData
{
GroupId = "1",
Features = 9,
Label = 1
};
var prediction = predictionEngine.Predict(testPage);
Console.WriteLine($"Predicted rating for: {prediction.Prediction}");

// New Page
testPage = new SearchData
{
GroupId = "2",
Features = 2,
Label = 9
};
prediction = predictionEngine.Predict(testPage);
Console.WriteLine($"Predicted: {prediction.Prediction}");

Console.WriteLine("Press any key to continue...");
Console.ReadKey();
}

private static void PrintMetrics(RankingMetrics metrics)
{
Console.WriteLine($"NormalizedDiscountedCumulativeGains: {metrics.NormalizedDiscountedCumulativeGains}");
Console.WriteLine($"DiscountedCumulativeGains: {metrics.DiscountedCumulativeGains}");

}

public static List<SearchData> GenerateData(int count=100)
{
var data = new List<SearchData>();
var rnd = new Random();
for(int i = 0; i < count; i++)
{
var newNode = new SearchData()
{
GroupId = rnd.Next(1, 5).ToString(), //nama yang di rating
Label = rnd.Next(0,4), //rating 0-4
Features = rnd.Next(1,100) //user atau identitas

};
data.Add(newNode);
}
return data;
}
}
public class SearchDataPrediction
{
[ColumnName("PredictedLabel")]public float Prediction;

public float Score { get; set; }
}
public class SearchData
{
[LoadColumn(0)]public string GroupId;

[LoadColumn(1)]public float Features;

[LoadColumn(2)]public float Label;
}


}

Nah kode diatas datanya kita generate dengan fungsi random. Nanti kita bisa prediksi jika user 1 dan groupid (search index) = 1 kira-kira berapa ratingnya ?  Kita bisa aplikasikan ke case study lain seperti rekomendasi buku, film, barang, dsb. Selamat berexperiment.

Salam Developer !!