public BICScoringFunction(BayesianNetwork network, RecordFile recordFile, LogLikelihoodCalculator llc, Constraints constraints)
{
this.network = network;
this.llc = llc;
this.constraints = constraints;
baseComplexityPenalty = Math.Log(recordFile.Size()) / 2;
}
public List<List<BitArray>> GetConsistentRecords(RecordFile recordFile)
{
List<List<BitArray>> consistentRecords = new List<List<BitArray>>();
for(int i = 0; i < Size(); i++)
{
consistentRecords.Add(new List<BitArray>());
int count = GetCardinality(i);
for(int value = 0; value < count; value++)
{
consistentRecords[i].Add(new BitArray(recordFile.Size()));
}
}
for (int index = 0; index < recordFile.Size(); index++)
{
Record record = recordFile.Records[index];
for (int variable = 0; variable < Size(); variable++)
{
string v = record[variable];
int value = Get(variable).ValueToIndex[v];
consistentRecords[variable][value].Set(index, true);
}
}
return consistentRecords;
}
private void Initialize(BayesianNetwork network, RecordFile recordFile)
{
this.network = network;
recordCount = recordFile.Size();
zero = new Varset(network.Size());
consistentRecords = network.GetConsistentRecords(recordFile);
}
public ADTree(int rMin, BayesianNetwork network, RecordFile recordFile)
{
this.rMin = rMin;
Initialize(network, recordFile);
CreateTree();
}
public void Execute(string[] args)
{
// 引数のチェック
if (args.Length != 2)
{
// エラーメッセージ表示
Console.WriteLine("[エラー] 引数の数が不正です");
Environment.Exit(0);
}
// オプションのチェック
// 暫定的に初期値を代入
int rMin = 5; // The minimum number of records in the AD-tree nodes.
char delimiter = ',';
bool hasHeader = true;
string sf = "BDeu";
int maxParents = 0;
string constraintsFile = "";
int runningTime = -1;
int threadCount = 20;
bool prune = true;
string inputFile = args[0];
string outputFile = args[1];
double equivarentSampleSize = 1;
// csvファイルの読み込み
RecordFile recordFile = new RecordFile();
recordFile.ReadRecord(inputFile, hasHeader, delimiter);
//recordFile.Print();
// BayesianNetworkの初期化
BayesianNetwork network = new BayesianNetwork(recordFile);
// AD-Treeの初期化
ADTree adTree = new ADTree(rMin, network, recordFile);
// scoring functionの設定
sf = sf.ToLower();
if (maxParents > network.Size() || maxParents < 1)
{
maxParents = network.Size() - 1;
}
if (sf == "bic")
{
int maxParentCount = (int)Math.Log(2 * recordFile.Size() / Math.Log(recordFile.Size()));
if(maxParentCount < maxParents)
{
maxParents = maxParentCount;
}
}
else if (sf == "fnml") { }
else if (sf == "bdeu") { }
else
{
throw new ArgumentException("Invalid scoring function. Options are: 'BIC', 'fNML' or 'BDeu'.");
}
// TODO constraintsの実装
Constraints constraints = null;
if (constraintsFile.Length > 0)
{
constraints = Constraints.ParseConstraints(constraintsFile, network);
}
ScoringFunction scoringFunction = null;
List<double> ilogi = LogLikelihoodCalculator.GetLogCache(recordFile.Size());
LogLikelihoodCalculator llc = new LogLikelihoodCalculator(adTree, network, ilogi);
// TODO regretの実装
//std::vector<std::vector<float>*>* regret = scoring::getRegretCache(recordFile.size(), network.getMaxCardinality());
if (sf == "bic")
{
scoringFunction = new BICScoringFunction(network, recordFile, llc, constraints);
}
else if (sf == "fnml")
{
//
}
else if (sf == "bdeu")
{
scoringFunction = new BDeuScoringFunction(equivarentSampleSize, network, adTree, constraints);
}
ScoreCalculator scoreCalculator = new ScoreCalculator(scoringFunction, maxParents, network.Size(), runningTime, constraints);
// TODO Parallelクラスによる並列処理
Parallel.For(0, threadCount, i =>
{
for (int variable = 0; variable < network.Size(); variable++)
{
if(variable % threadCount != i)
{
continue;
}
Console.WriteLine("Thread: " + i + ", Variable: " + variable + ", Time: " + DateTime.Now);
DoubleMap sc = new DoubleMap();
scoreCalculator.CalculateScores(variable, sc);
int size = sc.Count;
Console.WriteLine("Thread: " + i + ", Variable: " + variable + ", Size before pruning: " + size + ", Time: " + DateTime.Now);
if (prune)
{
scoreCalculator.Prune(sc);
int prunedSize = sc.Count;
Console.WriteLine("Thread: " + i + ", Variable: " + variable + ", Size after pruning: " + prunedSize + ", Time: " + DateTime.Now);
}
string varFilename = outputFile + "." + variable;
StreamWriter writer = new StreamWriter(varFilename, false);
Variable var = network.Get(variable);
writer.Write("VAR " + var.Name + "\n");
writer.Write("META arity=" + var.GetCardinality() + "\n");
writer.Write("META values=");
for (int k = 0; k < var.GetCardinality(); k++)
{
writer.Write(var.GetValue(k) + " ");
}
writer.Write("\n");
foreach (KeyValuePair<ulong, double> kvp in sc)
{
Varset parentSet = new Varset(kvp.Key);
double s = kvp.Value;
writer.Write(s + " ");
for (int p = 0; p < network.Size(); p++)
{
if (parentSet.Get(p))
{
writer.Write(network.Get(p).Name + " ");
}
}
writer.Write("\n");
}
writer.Write("\n");
writer.Close();
}
});
// concatenate all of the files together
StreamWriter outFile = new StreamWriter(outputFile, false); // 修正が必要かも
// first, the header information
string header = "META pss_version = 0.1\nMETA input_file=" + inputFile + "\nMETA num_records=" + recordFile.Size() + "\n";
header += "META parent_limit=" + maxParents + "\nMETA score_type=" + sf + "\nMETA ess=" + equivarentSampleSize + "\n\n";
outFile.Write(header);
for (int variable = 0; variable < network.Size(); variable++)
{
string varFilename = outputFile + "." + variable;
StreamReader reader = new StreamReader(varFilename);
outFile.Write(reader.ReadToEnd());
reader.Close();
}
outFile.Close();
}
public void AddValues(RecordFile recordFile)
{
recordFile.Records.ForEach(line => AddValue(line[index]));
}
private void Initialize(RecordFile recordFile)
{
// 変数名の設定
for (int i = 0; i < recordFile.Header.Count; i++)
{
Variable v = new Variable(this, i);
if (recordFile.HasHeader)
{
v.Name = recordFile.Header[i];
}
else
{
string variableName = "Variable_" + i;
v.Name = variableName;
}
nameToIndex[v.Name] = i;
variables.Add(v);
}
// 値をセットする
variables.ForEach(variable => variable.AddValues(recordFile));
}
public BayesianNetwork(RecordFile recordFile)
{
Initialize(recordFile);
}