summaryrefslogtreecommitdiff
path: root/lib/validation-statistics.rb
blob: 2dd9c7aefc6ee6bec899cf38321259ce0c77fd86 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
module OpenTox
  module Validation
    # Statistical evaluation of classification validations
    module ClassificationStatistics

      # Get statistics
      # @return [Hash]
      def statistics 
        self.accept_values = model.prediction_feature.accept_values
        self.confusion_matrix = {:all => Array.new(accept_values.size){Array.new(accept_values.size,0)}, :confidence_high => Array.new(accept_values.size){Array.new(accept_values.size,0)}, :confidence_low => Array.new(accept_values.size){Array.new(accept_values.size,0)}}
        self.nr_predictions = {:all => 0,:confidence_high => 0,:confidence_low => 0}
        predictions.each do |cid,pred|
          # TODO: use measured majority class or all measurements??
          if pred[:measurements].uniq.size == 1 and pred[:probabilities]
            m = pred[:measurements].first
            if pred[:value] == m
              accept_values.each_with_index do |v,i|
                if pred[:value] == v
                  confusion_matrix[:all][i][i] += 1
                  self.nr_predictions[:all] += 1
                  if pred[:confidence].match(/Similar/i)
                    confusion_matrix[:confidence_high][i][i] += 1
                    self.nr_predictions[:confidence_high] += 1
                  elsif pred[:confidence].match(/Low/i)
                    confusion_matrix[:confidence_low][i][i] += 1
                    self.nr_predictions[:confidence_low] += 1
                  end
                end
              end
            elsif pred[:value] != m
              accept_values.each_with_index do |v,i|
                if pred[:value] == v
                  confusion_matrix[:all][i][(i+1)%2] += 1
                  self.nr_predictions[:all] += 1
                  if pred[:confidence].match(/Similar/i)
                    confusion_matrix[:confidence_high][i][(i+1)%2] += 1
                    self.nr_predictions[:confidence_high] += 1
                  elsif pred[:confidence].match(/Low/i)
                    confusion_matrix[:confidence_low][i][(i+1)%2] += 1
                    self.nr_predictions[:confidence_low] += 1
                  end
                end
              end
            end
          end
        end

        self.true_rate = {:all => {}, :confidence_high => {}, :confidence_low => {}}
        self.predictivity = {:all => {}, :confidence_high => {}, :confidence_low => {}}
        accept_values.each_with_index do |v,i|
          [:all,:confidence_high,:confidence_low].each do |a|
            self.true_rate[a][v] = confusion_matrix[a][i][i]/confusion_matrix[a][i].reduce(:+).to_f
            self.predictivity[a][v] = confusion_matrix[a][i][i]/confusion_matrix[a].collect{|n| n[i]}.reduce(:+).to_f
          end
        end
        self.accuracy = {}
        [:all,:confidence_high,:confidence_low].each do |a|
          self.accuracy[a] = (confusion_matrix[a][0][0]+confusion_matrix[a][1][1])/nr_predictions[a].to_f
        end
        $logger.debug "Accuracy #{accuracy}"
        $logger.debug "Nr Predictions #{nr_predictions}"
        save
        {
          :accept_values => accept_values,
          :confusion_matrix => confusion_matrix,
          :accuracy => accuracy,
          :true_rate => self.true_rate,
          :predictivity => self.predictivity,
          :nr_predictions => nr_predictions,
        }
      end

      # Plot accuracy vs prediction probability
      # @param [String,nil] format
      # @return [Blob]
      def probability_plot format: "pdf"
        #unless probability_plot_id

          #tmpdir = File.join(ENV["HOME"], "tmp")
          tmpdir = "/tmp"
          #p tmpdir
          FileUtils.mkdir_p tmpdir
          tmpfile = File.join(tmpdir,"#{id.to_s}_probability.#{format}")
          accuracies = []
          probabilities = []
          correct_predictions = 0
          incorrect_predictions = 0
          pp = []
          predictions.values.select{|p| p["probabilities"]}.compact.each do |p|
            p["measurements"].each do |m|
              pp << [ p["probabilities"][p["value"]], p["value"] == m ]
            end
          end
          pp.sort_by!{|p| 1-p.first}
          pp.each do |p|
            p[1] ? correct_predictions += 1 : incorrect_predictions += 1
            accuracies << correct_predictions/(correct_predictions+incorrect_predictions).to_f
            probabilities << p[0]
          end
          R.assign "accuracy", accuracies
          R.assign "probability", probabilities
          R.eval "image = qplot(probability,accuracy)+ylab('Accumulated accuracy')+xlab('Prediction probability')+ylim(c(0,1))+scale_x_reverse()+geom_line()"
          R.eval "ggsave(file='#{tmpfile}', plot=image)"
          file = Mongo::Grid::File.new(File.read(tmpfile), :filename => "#{self.id.to_s}_probability_plot.svg")
          plot_id = $gridfs.insert_one(file)
          update(:probability_plot_id => plot_id)
        #end
        $gridfs.find_one(_id: probability_plot_id).data
      end
    end

    # Statistical evaluation of regression validations
    module RegressionStatistics

      attr_accessor :x, :y

      # Get statistics
      # @return [Hash]
      def statistics
        self.warnings = []
        self.rmse = {:all =>0,:confidence_high => 0,:confidence_low => 0}
        self.r_squared  = {:all =>0,:confidence_high => 0,:confidence_low => 0}
        self.mae = {:all =>0,:confidence_high => 0,:confidence_low => 0}
        self.within_prediction_interval = {:all =>0,:confidence_high => 0,:confidence_low => 0}
        self.out_of_prediction_interval = {:all =>0,:confidence_high => 0,:confidence_low => 0}
        @x = {:all => [],:confidence_high => [],:confidence_low => []}
        @y = {:all => [],:confidence_high => [],:confidence_low => []}
        self.nr_predictions = {:all =>0,:confidence_high => 0,:confidence_low => 0}
        predictions.each do |cid,pred|
          !if pred[:value] and pred[:measurements] and !pred[:measurements].empty?
            insert_prediction pred, :all
            if pred[:confidence].match(/Similar/i)
              insert_prediction pred, :confidence_high
            elsif pred[:confidence].match(/Low/i)
              insert_prediction pred, :confidence_low
            end
          else
            trd_id = model.training_dataset_id
            smiles = Compound.find(cid).smiles
            self.warnings << "No training activities for #{smiles} in training dataset #{trd_id}."
            $logger.debug "No training activities for #{smiles} in training dataset #{trd_id}."
          end
        end
        [:all,:confidence_high,:confidence_low].each do |a|
          if @x[a].size > 2
            R.assign "measurement", @x[a]
            R.assign "prediction", @y[a]
            R.eval "r <- cor(measurement,prediction,use='pairwise')"
            self.r_squared[a] = R.eval("r").to_ruby**2
          else
            self.r_squared[a] = 0
          end
          if self.nr_predictions[a] > 0
            self.mae[a] = self.mae[a]/self.nr_predictions[a]
            self.rmse[a] = Math.sqrt(self.rmse[a]/self.nr_predictions[a])
          else
            self.mae[a] = nil
            self.rmse[a] = nil
          end
        end
        $logger.debug "R^2 #{r_squared}"
        $logger.debug "RMSE #{rmse}"
        $logger.debug "MAE #{mae}"
        $logger.debug "Nr predictions #{nr_predictions}"
        $logger.debug "#{within_prediction_interval} measurements within prediction interval"
        save
        {
          :mae => mae,
          :rmse => rmse,
          :r_squared => r_squared,
          :within_prediction_interval => self.within_prediction_interval,
          :out_of_prediction_interval => out_of_prediction_interval,
          :nr_predictions => nr_predictions,
        }
      end

      # Plot predicted vs measured values
      # @param [String,nil] format
      # @return [Blob]
      def correlation_plot format: "png"
        unless correlation_plot_id
          tmpfile = "/tmp/#{id.to_s}_correlation.#{format}"
          x = []
          y = []
          feature = Feature.find(predictions.first.last["prediction_feature_id"])
          predictions.each do |sid,p|
            x << p["measurements"].median
            y << p["value"]
          end
          R.assign "measurement", x
          R.assign "prediction", y
          R.eval "all = c(measurement,prediction)"
          R.eval "range = c(min(all), max(all))"
          if feature.name.match /Net cell association/ # ad hoc fix for awkward units
            title = "log2(Net cell association [mL/ug(Mg)])"
          else
            title = feature.name
            title += "-log10(#{feature.unit})" if feature.unit and !feature.unit.blank?
          end
          R.eval "image = qplot(prediction,measurement,main='#{title}',xlab='Prediction',ylab='Measurement',asp=1,xlim=range, ylim=range)"
          R.eval "image = image + geom_abline(intercept=0, slope=1)"
          R.eval "ggsave(file='#{tmpfile}', plot=image)"
          file = Mongo::Grid::File.new(File.read(tmpfile), :filename => "#{id.to_s}_correlation_plot.#{format}")
          plot_id = $gridfs.insert_one(file)
          update(:correlation_plot_id => plot_id)
        end
        $gridfs.find_one(_id: correlation_plot_id).data
      end

      # Get predictions with measurements outside of the prediction interval
      # @return [Hash]
      def worst_predictions
        worst_predictions = predictions.select do |sid,p|
          p["prediction_interval"] and p["value"] and (p["measurements"].max < p["prediction_interval"][0] or p["measurements"].min > p["prediction_interval"][1])
        end.compact.to_h
        worst_predictions.each do |sid,p|
          p["error"] = (p["value"] - p["measurements"].median).abs
          if p["measurements"].max < p["prediction_interval"][0]
            p["distance_prediction_interval"] = (p["measurements"].max - p["prediction_interval"][0]).abs
          elsif p["measurements"].min > p["prediction_interval"][1]
            p["distance_prediction_interval"] = (p["measurements"].min - p["prediction_interval"][1]).abs
          end
        end
        worst_predictions.sort_by{|sid,p| p["distance_prediction_interval"] }.to_h
      end

      private

      def insert_prediction prediction, type
        self.nr_predictions[type] +=1
        @x[type] << prediction[:measurements].median
        @y[type] << prediction[:value]
        error = prediction[:value]-prediction[:measurements].median
        self.rmse[type] += error**2
        self.mae[type] += error.abs
        if prediction[:prediction_interval]
          if prediction[:measurements].median >= prediction[:prediction_interval][0] and prediction[:measurements].median <= prediction[:prediction_interval][1]
            self.within_prediction_interval[type] += 1
          else
            self.out_of_prediction_interval[type] += 1
          end
        end
      end
    end
  end
end