public Dictionary <string, string> ExportSettings(bool suppressWarnings)
{
Dictionary <string, string> SettingsDict = new Dictionary <string, string>();
SettingsDict.Add("RawFreqs", RawFreqs.ToString());
//if we have a dictionary loaded, then we'll save it
if (!String.IsNullOrEmpty(DictionaryLocation))
{
SettingsDict.Add("DictionaryLocation", DictionaryLocation.ToString());
SettingsDict.Add("DictionaryContents", UserLoadedDictionary.DictionaryRawText);
SettingsDict.Add("IncludeStDev", IncludeStDevs.ToString());
SettingsDict.Add("RoundLength", RoundValuesToNDecimals.ToString());
}
return(SettingsDict);
}
//one of the few plugins thus far where I'm actually using a constructor
//might not be the most efficient way to handle this (especially at runtime)
//but I don't suspect that it'll be too bad.
//public ExamineDictWords()
//{
// DictionaryList = new List<DictionaryMetaObject>();
// ListOfBuiltInDictionaries = new HashSet<string>();
// foreach(DictionaryMetaObject dict in DictionaryList)
// {
// ListOfBuiltInDictionaries.Add(dict.DictionaryName);
// }
//}
public Payload FinishUp(Payload Input)
{
Payload OutputData = new Payload();
OutputData.FileID = "";
string StringOutputFormatParameter = "N" + RoundValuesToNDecimals.ToString();
#region Cronbach's Alphas
//first thing's first: we have to figure out which words constitute each category so that we can calculate Σ[s2i]
Dictionary <int, List <string> > CatWordMap = new Dictionary <int, List <String> >();
for (int i = 0; i < UserLoadedDictionary.DictData.NumCats; i++)
{
CatWordMap.Add(i, new List <string>());
}
#region Figure out which words belong to which category
//iterate over n-grams, starting with the largest possible n-gram (derived from the user's dictionary file)
for (int NumberOfWords = UserLoadedDictionary.DictData.MaxWords; NumberOfWords > 0; NumberOfWords--)
{
if (UserLoadedDictionary.DictData.FullDictionary["Standards"].ContainsKey(NumberOfWords))
{
foreach (string term in UserLoadedDictionary.DictData.FullDictionary["Standards"][NumberOfWords].Keys)
{
for (int wordCatCount = 0; wordCatCount < UserLoadedDictionary.DictData.FullDictionary["Standards"][NumberOfWords][term].Length; wordCatCount++)
{
int outputCatMap = OutputDataMap[UserLoadedDictionary.DictData.FullDictionary["Standards"][NumberOfWords][term][wordCatCount]];
CatWordMap[outputCatMap].Add(term);
}
}
}
if (UserLoadedDictionary.DictData.FullDictionary["Wildcards"].ContainsKey(NumberOfWords))
{
foreach (string term in UserLoadedDictionary.DictData.FullDictionary["Wildcards"][NumberOfWords].Keys)
{
for (int wordCatCount = 0; wordCatCount < UserLoadedDictionary.DictData.FullDictionary["Wildcards"][NumberOfWords][term].Length; wordCatCount++)
{
int outputCatMap = OutputDataMap[UserLoadedDictionary.DictData.FullDictionary["Wildcards"][NumberOfWords][term][wordCatCount]];
CatWordMap[outputCatMap].Add(term);
}
}
}
}
#endregion
#region Raw Cronbach
OutputData.SegmentNumber.Add(0); //used as the entry number
OutputData.SegmentID.Add("Cronbach's Alpha (Raw)"); //the dictionary entry
string[] OutputArray_Cronbach = new string[UserLoadedDictionary.DictData.NumCats];
//now, we go through each category and calculate the *raw* cronbach's alpha
for (int i = 0; i < UserLoadedDictionary.DictData.NumCats; i++)
{
//this gets us to sum of variances for the category's constituent items
double itemVarianceSum = 0;
string[] catWordList = CatWordMap[i].ToArray();
double k = (double)catWordList.Length;
for (int j = 0; j < catWordList.Length; j++)
{
double itemVariance = TermVariancesRaw[catWordList[j]]["S"] / (TotalNumberOfDocs["Docs"] - 1);
if (itemVariance > 0)
{
itemVarianceSum += itemVariance;
}
else
{
k -= 1;
}
}
double totalVariance = CategoryVariancesRaw[i]["S"] / (TotalNumberOfDocs["Docs"] - 1);
//https://data.library.virginia.edu/using-and-interpreting-cronbachs-alpha/
//double CronbachRaw = (k / (k - 1)) * ((totalVariance - itemVarianceSum) / totalVariance);
double CronbachRaw = (k / (k - 1)) * (1 - (itemVarianceSum / totalVariance));
if (!Double.IsNaN(CronbachRaw) && !Double.IsInfinity(CronbachRaw))
{
OutputArray_Cronbach[i] = Math.Round(CronbachRaw, RoundValuesToNDecimals, MidpointRounding.AwayFromZero).ToString(StringOutputFormatParameter);
}
else
{
OutputArray_Cronbach[i] = "N/A";
}
}
OutputData.StringArrayList.Add(OutputArray_Cronbach);
#endregion
#region OneHot Cronbach
OutputData.SegmentNumber.Add(0); //used as the entry number
OutputData.SegmentID.Add("Kuder–Richardson Formula 20 (KR-20)"); //the dictionary entry
OutputArray_Cronbach = new string[UserLoadedDictionary.DictData.NumCats];
//now, we go through each category and calculate the *raw* cronbach's alpha
for (int i = 0; i < UserLoadedDictionary.DictData.NumCats; i++)
{
//this gets us to sum of variances for the category's constituent items
double itemVarianceSum = 0;
string[] catWordList = CatWordMap[i].ToArray();
double k = (double)catWordList.Length;
for (int j = 0; j < catWordList.Length; j++)
{
double itemVariance = TermVariancesOneHot[catWordList[j]]["S"] / (TotalNumberOfDocs["Docs"] - 1);
if (itemVariance > 0)
{
itemVarianceSum += itemVariance;
}
else
{
k -= 1;
}
}
double totalVariance = CategoryVariancesOneHot[i]["S"] / (TotalNumberOfDocs["Docs"] - 1);
//double CronbachOneHot = (k / (k - 1)) * ((totalVariance - itemVarianceSum) / totalVariance);
double CronbachOneHot = (k / (k - 1)) * (1 - (itemVarianceSum / totalVariance));
if (!Double.IsNaN(CronbachOneHot) && !Double.IsInfinity(CronbachOneHot))
{
OutputArray_Cronbach[i] = Math.Round(CronbachOneHot, RoundValuesToNDecimals, MidpointRounding.AwayFromZero).ToString(StringOutputFormatParameter);
}
else
{
OutputArray_Cronbach[i] = "N/A";
}
}
OutputData.StringArrayList.Add(OutputArray_Cronbach);
#endregion
#endregion
#region this is where we calculate the avg percentages for each word
for (int i = 0; i < UserLoadedDictionary.DictData.AllEntries.Count; i++)
{
OutputData.SegmentNumber.Add((ulong)(i + 1)); //used as the entry number
OutputData.SegmentID.Add(UserLoadedDictionary.DictData.AllEntries[i]); //the dictionary entry
string[] OutputArray = new string[UserLoadedDictionary.DictData.NumCats];
for (int j = 0; j < UserLoadedDictionary.DictData.NumCats; j++)
{
OutputArray[j] = "";
}
for (int j = 0; j < UserLoadedDictionary.DictData.NumCats; j++)
{
//if we know that the mean is zero, then we just skip on to the next one
if (EntryFreqTracker_Long[UserLoadedDictionary.DictData.AllEntries[i]][j] == 0)
{
continue;
}
if (RawFreqs)
{
OutputArray[j] = Math.Round((EntryFreqTracker_Long[UserLoadedDictionary.DictData.AllEntries[i]][j] / (double)TotalNumberOfDocs["Docs"]), RoundValuesToNDecimals, MidpointRounding.AwayFromZero).ToString(StringOutputFormatParameter);
}
else
{
OutputArray[j] = Math.Round(((EntryFreqTracker_Double[UserLoadedDictionary.DictData.AllEntries[i]][j] / TotalNumberOfDocs["Docs"]) * 100), RoundValuesToNDecimals, MidpointRounding.AwayFromZero).ToString(StringOutputFormatParameter);
}
//calculate the standard deviation
if (IncludeStDevs)
{
OutputArray[j] += " (" + Math.Round(Math.Sqrt(TermVariancesRaw[UserLoadedDictionary.DictData.AllEntries[i]]["S"]), RoundValuesToNDecimals, MidpointRounding.AwayFromZero).ToString(StringOutputFormatParameter) + ")";
}
}
OutputData.StringArrayList.Add(OutputArray);
}
#endregion
EntryFreqTracker_Long = new ConcurrentDictionary <string, ulong[]>();
EntryFreqTracker_Double = new ConcurrentDictionary <string, double[]>();
return(OutputData);
}
