public Dictionary <string, double> GetWeights(List <Sample> samples)
{
int nAttributes = 13;
int nSamples = samples.Count;
double[,] tsData = new double[nSamples, nAttributes];
double[] resultData = new double[nSamples];
for (int i = 0; i < samples.Count; i++)
{
tsData[i, 0] = samples[i].Precipitation;
tsData[i, 1] = samples[i].SoilWatFieldCapacity;
tsData[i, 2] = samples[i].Swawp;
tsData[i, 3] = samples[i].SoilWaterStorage;
tsData[i, 4] = samples[i].IrrigationEfficiency;
tsData[i, 5] = samples[i].MaxYield;
tsData[i, 6] = samples[i].GrossIrrig;
tsData[i, 7] = samples[i].InputCosts;
tsData[i, 8] = samples[i].OperationCosts;
tsData[i, 9] = samples[i].MiscCosts;
tsData[i, 10] = samples[i].IrrigCosts;
tsData[i, 11] = samples[i].MiscReturns;
tsData[i, 6] = samples[i].TotalIrrig;
tsData[i, 13] = samples[i].PriceUnit;
resultData[i] = samples[i].Yield;
}
double[] weights = null;
int fitResult = 0;
alglib.lsfit.lsfitreport rep = new alglib.lsfit.lsfitreport();
alglib.lsfit.lsfitlinear(resultData, tsData, nSamples, nAttributes, ref fitResult, ref weights, rep);
Dictionary <string, double> labelsAndWeights = new Dictionary <string, double>();
labelsAndWeights.Add("Precipitation", weights[0]);
labelsAndWeights.Add("SoilWatFieldCapacity", weights[1]);
labelsAndWeights.Add("Swawp", weights[2]);
labelsAndWeights.Add("SoilWaterStorage", weights[2]);
labelsAndWeights.Add("IrrigationEfficiency", weights[3]);
labelsAndWeights.Add("MaxYield", weights[4]);
labelsAndWeights.Add("GrossIrrig", weights[5]);
labelsAndWeights.Add("InputCosts", weights[6]);
labelsAndWeights.Add("OperationCosts", weights[7]);
labelsAndWeights.Add("MiscCosts", weights[8]);
labelsAndWeights.Add("IrrigCosts", weights[9]);
labelsAndWeights.Add("MiscReturns", weights[10]);
labelsAndWeights.Add("TotalIrrig", weights[11]);
labelsAndWeights.Add("PriceUnit", weights[12]);
return(labelsAndWeights);
}
public static double PredictStockPrice(string SearchTerm, string StockCode, DateTime FromDate, DateTime ToDate)
{
Console.WriteLine("Retrieving historic data ({0} days available)", ToDate.Subtract(FromDate).Days);
var samples = RetrieveEnties(SearchTerm, StockCode, FromDate, ToDate);
Console.WriteLine("Calculating multiple regression");
int nAttributes = 3;
int nSamples = samples.Count;
double[,] tsData = new double[nSamples, nAttributes];
double[] resultData = new double[nSamples];
for (int i = 0; i < samples.Count; i++)
{
tsData[i, 0] = samples[i].FollowerCount;
tsData[i, 1] = samples[i].MaxTweetDensity;
tsData[i, 2] = samples[i].PositiveSentiment;
resultData[i] = samples[i].ClosingPrice;
}
double[] weights = null;
int fitResult = 0;
alglib.lsfit.lsfitreport rep = new alglib.lsfit.lsfitreport();
alglib.lsfit.lsfitlinear(resultData, tsData, nSamples, nAttributes, ref fitResult, ref weights, rep);
Console.WriteLine("Retrieving latest tweets");
var tweets = DataRetrieve.SearchTwitter(SearchTerm);
var dayData = from t in tweets
where t.CreatedAt.Date > DateTime.Today.AddDays(-_dayThreshold) &&
t.CreatedAt.Date <= DateTime.Today
select t;
var followerCount = (from row in dayData
select row.FollowersCount + 1)
.Sum();
var positive = (from row in dayData
select row.Positive)
.Sum() / dayData.Count();
var maxDensity = 0d;
var time = DateTime.Now.AddHours(7);
for (int i = 0; i <= 24 * _dayThreshold; i++)
{
var select = (from row in dayData
where row.CreatedAt >= time.AddDays(-_dayThreshold).AddHours(i) &&
row.CreatedAt < time.AddDays(-_dayThreshold).AddHours(i + 1)
select row.Positive);
var density = select.Sum();
maxDensity = Math.Max(maxDensity, density);
}
var predictedPrice = weights[0] * followerCount +
weights[1] * maxDensity +
weights[2] * positive;
return(predictedPrice);
}
/// <summary>
///
/// </summary>
/// <param name="tables"></param>diningArea.tables
/// <param name="customers"></param>diningArea.customers
/// <param name="orders"></param>diningArea.orders
/// <param name="finishedOrders"></param>training data
/// <param name="allItem"></param>
//make sure the itemId is continuous
public void PredictWaitingTimeReg(List <Table> tables, List <Customer> customers, List <Order> orders, List <Order> finishedOrders, List <Item> allItem)
{
//HashSet<Item> hs = new HashSet<Item>(allItem);//去重
int nAttributes = allItem.Count + 1;
int nSamples = finishedOrders.Count;
double[,] tsData = new double[nSamples, nAttributes];
double[] resultData = new double[nSamples];
for (int i = 0; i < nSamples; i++)
{
for (int j = 0; j < nAttributes; j++)
{
tsData[i, j] = 0;
}
tsData[i, 0] = finishedOrders[i].NumberofPeople;
foreach (var item in finishedOrders[i].Items)
{
tsData[i, item.ItemId] = item.ItemAmount;
}
TimeSpan ts = finishedOrders[i].FinishTime - finishedOrders[i].StartTime;
resultData[i] = ts.TotalMinutes;
}
double[] weights = null;
int fitResult = 0;
alglib.lsfit.lsfitreport rep = new alglib.lsfit.lsfitreport();
alglib.lsfit.lsfitlinear(resultData, tsData, nSamples, nAttributes, ref fitResult, ref weights, rep);
//about weight: weights[0] -> NumberofPeople,weights[1-allItem.Count]->ItemId,weights[allItem.Count+1 - nAttributes]->ItemAmount
foreach (var t in tables)
{
if (t.TableStatus == TableStatus.Active)
{
t.CountDown = System.Data.SqlTypes.SqlDateTime.MinValue.Value;
}
else if (t.TableStatus == TableStatus.Cleaning)
{
//t.CountDown = DateTime.Now.AddMinutes(5);
//this operation will be done when set tablestatus to cleaning.
}
else if (t.TableStatus == TableStatus.Reserved || t.TableStatus == TableStatus.Breakdown)
{
t.CountDown = System.Data.SqlTypes.SqlDateTime.MaxValue.Value;
}
else if (fitResult == 1 && t.OrderId > 0)//solved
{
Order currentOrder = orders.Find(x => x.OrderId == t.OrderId);
double waitTime = weights[0] * currentOrder.NumberofPeople;
foreach (var item in currentOrder.Items)
{
waitTime += weights[item.ItemId] * item.ItemAmount;
}
t.CountDown = currentOrder.StartTime.AddMinutes(waitTime);
}
}
//Dictionary<int, double> labelsAndWeights = new Dictionary<int, double>();
//labelsAndWeights.Add(0, weights[0]);
//labelsAndWeights.Add(1, weights[1]);
//labelsAndWeights.Add(2, weights[2]);
}