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
|
import { NgIf } from "@angular/common";
import { FolderFile } from "./FolderFile";
export default class Model extends FolderFile {
constructor(
name: string = 'Novi model',
public description: string = '',
dateCreated: Date = new Date(),
lastUpdated: Date = new Date(),
//public experimentId: string = '',
// Neural net training settings
public type: ProblemType = ProblemType.Regression,
public optimizer: Optimizer = Optimizer.Adam,
public lossFunction: LossFunction = LossFunctionRegression[0],
public inputNeurons: number = 1,
public hiddenLayers: number = 1,
public batchSize: BatchSize = BatchSize.O3,
public outputLayerActivationFunction: ActivationFunction = ActivationFunction.Sigmoid,
public uploaderId: string = '',
public metrics: string[] = [], // TODO add to add-model form
public epochs: number = 5, // TODO add to add-model form
public inputColNum: number = 5,
public learningRate: LearningRate = LearningRate.LR1,
public layers: Layer[] = [new Layer()],
// Test set settings
public randomOrder: boolean = true,
public randomTestSet: boolean = true,
public randomTestSetDistribution: number = 0.1, //0.1-0.9 (10% - 90%) JESTE OVDE ZAKUCANO 10, AL POSLATO JE KAO 0.1 BACK-U
public isPublic: boolean = false,
public accessibleByLink: boolean = false
) {
super(name, dateCreated, lastUpdated);
}
}
export class Layer {
constructor(
public layerNumber: number = 0,
public activationFunction: ActivationFunction = ActivationFunction.Sigmoid,
public neurons: number = 3,
public regularisation: Regularisation = Regularisation.L1,
public regularisationRate: RegularisationRate = RegularisationRate.RR1,
) { }
}
export enum LearningRate {
LR1 = '0.00001',
LR2 = '0.0001',
LR3 = '0.001',
LR4 = '0.003',
LR5 = '0.01',
LR6 = '0.03',
LR7 = '0.1',
LR8 = '0.3',
LR9 = '1',
LR10 = '3',
LR11 = '10',
}
export enum Regularisation {
L1 = 'l1',
L2 = 'l2'
}
export enum RegularisationRate {
RR1 = '0',
RR2 = '0.001',
RR3 = '0.003',
RR4 = '0.01',
RR5 = '0.03',
RR6 = '0.1',
RR7 = '0.3',
RR8 = '1',
RR9 = '3',
RR10 = '10',
}
export enum ProblemType {
Regression = 'regresioni',
BinaryClassification = 'binarni-klasifikacioni',
MultiClassification = 'multi-klasifikacioni'
}
// replaceMissing srednja vrednost mean, median, najcesca vrednost (mode)
// removeOutliers
export enum ActivationFunction {
// linear
Binary_Step = 'binaryStep',
// non-linear
Leaky_Relu = 'leakyRelu',
Parameterised_Relu = 'parameterisedRelu',
Exponential_Linear_Unit = 'exponentialLinearUnit',
Swish = 'swish',
//hiddenLayers
Relu = 'relu',
Sigmoid = 'sigmoid',
Tanh = 'tanh',
//outputLayer
Linear = 'linear',
//Sigmoid='sigmoid',
Softmax = 'softmax',
}
/*
export enum ActivationFunctionHiddenLayer
{
Relu='relu',
Sigmoid='sigmoid',
Tanh='tanh'
}
export enum ActivationFunctionOutputLayer
{
Linear = 'linear',
Sigmoid='sigmoid',
Softmax='softmax'
}
*/
export enum LossFunction {
// binary classification loss functions
BinaryCrossEntropy = 'binary_crossentropy',
SquaredHingeLoss = 'squared_hinge_loss',
HingeLoss = 'hinge_loss',
// multi-class classification loss functions
CategoricalCrossEntropy = 'categorical_crossentropy',
SparseCategoricalCrossEntropy = 'sparse_categorical_crossentropy',
KLDivergence = 'kullback_leibler_divergence',
// regression loss functions
MeanAbsoluteError = 'mean_absolute_error',
MeanSquaredError = 'mean_squared_error',
MeanSquaredLogarithmicError = 'mean_squared_logarithmic_error',
HuberLoss = 'Huber'
}
export const LossFunctionRegression = [LossFunction.MeanAbsoluteError, LossFunction.MeanSquaredError, LossFunction.MeanSquaredLogarithmicError]
export const LossFunctionBinaryClassification = [LossFunction.BinaryCrossEntropy, LossFunction.SquaredHingeLoss, LossFunction.HingeLoss]
export const LossFunctionMultiClassification = [LossFunction.CategoricalCrossEntropy, LossFunction.SparseCategoricalCrossEntropy, LossFunction.KLDivergence]
export enum Optimizer {
Adam = 'Adam',
Adadelta = 'Adadelta',
Adagrad = 'Adagrad',
Ftrl = 'Ftrl',
Nadam = 'Nadam',
SGD = 'SGD',
SGDMomentum = 'SGDMomentum',
RMSprop = 'RMSprop'
}
export enum NullValueOptions {
DeleteRows = 'delete_rows',
DeleteColumns = 'delete_columns',
Replace = 'replace'
}
export enum ReplaceWith {
None = 'Popuni...',
Mean = 'Srednja vrednost',
Median = 'Medijana'
}
export class NullValReplacer {
"column": string;
"option": NullValueOptions;
"value": string;
}
export enum Metrics {
MSE = 'mse',
MAE = 'mae',
RMSE = 'rmse'
}
export enum MetricsRegression {
Mse = 'mse',
Mae = 'mae',
Mape = 'mape',
Msle = 'msle',
CosineProximity = 'cosine'
}
export enum MetricsBinaryClassification {
Accuracy = 'binary_accuracy',
Auc = "AUC",
Precision = 'precision_score',
Recall = 'recall_score',
F1 = 'f1_score',
}
export enum MetricsMultiClassification {
Accuracy = 'categorical_accuracy',
Auc = "AUC",
Precision = 'precision_score',
Recall = 'recall_score',
F1 = 'f1_score',
}
export enum BatchSize {
O1 = '2',
O2 = '4',
O3 = '8',
O4 = '16',
O5 = '32',
O6 = '64',
O7 = '128',
O8 = '256',
O9 = '512',
O10 = '1024'
}
|
