# Neural Network Case Study

0 Comments / 1 / By

In this section we’ll walk through a complete implementation of a toy Neural Network in 2 dimensions. We’ll first implement a simple linear classifier and then extend the code to a 2-layer Neural Network. As we’ll see, this extension is surprisingly simple and very few changes are necessary.

## Generating some data

Lets generate a classification dataset that is not easily linearly separable. Our favorite example is the spiral dataset, which can be generated as follows:

The toy spiral data consists of three classes (blue, red, yellow) that are not linearly separable.

Normally we would want to preprocess the dataset so that each feature has zero mean and unit standard deviation, but in this case the features are already in a nice range from -1 to 1, so we skip this step.

## Training a Softmax Linear Classifier

### Initialize the parameters

Lets first train a Softmax classifier on this classification dataset. As we saw in the previous sections, the Softmax classifier has a linear score function and uses the cross-entropy loss. The parameters of the linear classifier consist of a weight matrix and a bias vector for each class. Lets first initialize these parameters to be random numbers:

Recall that we is the dimensionality and is the number of classes.

### Compute the class scores

Since this is a linear classifier, we can compute all class scores very simply in parallel with a single matrix multiplication:

In this example we have 300 2-D points, so after this multiplication the array will have size [300 x 3], where each row gives the class scores corresponding to the 3 classes (blue, red, yellow).

### Compute the loss

The second key ingredient we need is a loss function, which is a differentiable objective that quantifies our unhappiness with the computed class scores. Intuitively, we want the correct class to have a higher score than the other classes. When this is the case, the loss should be low and otherwise the loss should be high. There are many ways to quantify this intuition, but in this example lets use the cross-entropy loss that is associated with the Softmax classifier. Recall that if $$f$$ is the array of class scores for a single example (e.g. array of 3 numbers here), then the Softmax classifier computes the loss for that example as:

We can see that the Softmax classifier interprets every element of $$f$$ as holding the (unnormalized) log probabilities of the three classes. We exponentiate these to get (unnormalized) probabilities, and then normalize them to get probabilites. Therefore, the expression inside the log is the normalized probability of the correct class. Note how this expression works: this quantity is always between 0 and 1. When the probability of the correct class is very small (near 0), the loss will go towards (positive) infinity. Conversely, when the correct class probability goes towards 1, the loss will go towards zero because $$log(1) = 0$$. Hence, the expression for $$L_i$$ is low when the correct class probability is high, and it’s very high when it is low.

Recall also that the full Softmax classifier loss is then defined as the average cross-entropy loss over the training examples and the regularization:

Given the array of we’ve computed above, we can compute the loss. First, the way to obtain the probabilities is straight forward:

We now have an array of size [300 x 3], where each row now contains the class probabilities. In particular, since we’ve normalized them every row now sums to one. We can now query for the log probabilities assigned to the correct classes in each example:

The array is a 1D array of just the probabilities assigned to the correct classes for each example. The full loss is then the average of these log probabilities and the regularization loss:

In this code, the regularization strength $$\lambda$$ is stored inside the . The convenience factor of multiplying the regularization will become clear in a second. Evaluating this in the beginning (with random parameters) might give us , which is , since with small initial random weights all probabilities assigned to all classes are about one third. We now want to make the loss as low as possible, with as the absolute lower bound. But the lower the loss is, the higher are the probabilities assigned to the correct classes for all examples.

### Computing the Analytic Gradient with Backpropagation

We have a way of evaluating the loss, and now we have to minimize it. We’ll do so with gradient descent. That is, we start with random parameters (as shown above), and evaluate the gradient of the loss function with respect to the parameters, so that we know how we should change the parameters to decrease the loss. Lets introduce the intermediate variable $$p$$, which is a vector of the (normalized) probabilities. The loss for one example is:

We now wish to understand how the computed scores inside $$f$$ should change to decrease the loss $$L_i$$ that this example contributes to the full objective. In other words, we want to derive the gradient $$\partial L_i / \partial f_k$$. The loss $$L_i$$ is computed from $$p$$, which in turn depends on $$f$$. It’s a fun exercise to the reader to use the chain rule to derive the gradient, but it turns out to be extremely simple and interpretible in the end, after a lot of things cancel out:

Notice how elegant and simple this expression is. Suppose the probabilities we computed were , and that the correct class was the middle one (with probability 0.3). According to this derivation the gradient on the scores would be . Recalling what the interpretation of the gradient, we see that this result is highly intuitive: increasing the first or last element of the score vector (the scores of the incorrect classes) leads to an increased loss (due to the positive signs +0.2 and +0.5) - and increasing the loss is bad, as expected. However, increasing the score of the correct class has negative influence on the loss. The gradient of -0.7 is telling us that increasing the correct class score would lead to a decrease of the loss $$L_i$$, which makes sense.

All of this boils down to the following code. Recall that stores the probabilities of all classes (as rows) for each example. To get the gradient on the scores, which we call , we proceed as follows:

Lastly, we had that , so armed with the gradient on (stored in ), we can now backpropagate into and :

Where we see that we have backpropped through the matrix multiply operation, and also added the contribution from the regularization. Note that the regularization gradient has the very simple form since we used the constant for its loss contribution (i.e. $$\frac{d}{dw} ( \frac{1}{2} \lambda w^2) = \lambda w$$. This is a common convenience trick that simplifies the gradient expression.

### Performing a parameter update

Now that we’ve evaluated the gradient we know how every parameter influences the loss function. We will now perform a parameter update in the negative gradient direction to decrease the loss:

### Putting it all together: Training a Softmax Classifier

Putting all of this together, here is the full code for training a Softmax classifier with Gradient descent:

Running this prints the output:

We see that we’ve converged to something after about 190 iterations. We can evaluate the training set accuracy:

This prints 49%. Not very good at all, but also not surprising given that the dataset is constructed so it is not linearly separable. We can also plot the learned decision boundaries:

Linear classifier fails to learn the toy spiral dataset.

## Training a Neural Network

Clearly, a linear classifier is inadequate for this dataset and we would like to use a Neural Network. One additional hidden layer will suffice for this toy data. We will now need two sets of weights and biases (for the first and second layers):

The forward pass to compute scores now changes form:

Notice that the only change from before is one extra line of code, where we first compute the hidden layer representation and then the scores based on this hidden layer. Crucially, we’ve also added a non-linearity, which in this case is simple ReLU that thresholds the activations on the hidden layer at zero.

Everything else remains the same. We compute the loss based on the scores exactly as before, and get the gradient for the scores exactly as before. However, the way we backpropagate that gradient into the model parameters now changes form, of course. First lets backpropagate the second layer of the Neural Network. This looks identical to the code we had for the Softmax classifier, except we’re replacing (the raw data), with the variable ):

However, unlike before we are not yet done, because is itself a function of other parameters and the data! We need to continue backpropagation through this variable. Its gradient can be computed as:

Now we have the gradient on the outputs of the hidden layer. Next, we have to backpropagate the ReLU non-linearity. This turns out to be easy because ReLU during the backward pass is effectively a switch. Since $$r = max(0, x)$$, we have that $$\frac{dr}{dx} = 1(x > 0)$$. Combined with the chain rule, we see that the ReLU unit lets the gradient pass through unchanged if its input was greater than 0, but kills it if its input was less than zero during the forward pass. Hence, we can backpropagate the ReLU in place simply with:

And now we finally continue to the first layer weights and biases:

We’re done! We have the gradients and can perform the parameter update. Everything else remains unchanged. The full code looks very similar:

This prints:

The training accuracy is now:

Which prints 98%!. We can also visualize the decision boundaries:

Neural Network classifier crushes the spiral dataset.

## Summary

We’ve worked with a toy 2D dataset and trained both a linear network and a 2-layer Neural Network. We saw that the change from a linear classifier to a Neural Network involves very few changes in the code. The score function changes its form (1 line of code difference), and the backpropagation changes its form (we have to perform one more round of backprop through the hidden layer to the first layer of the network).

## Abstract

Background Artificial neural networks (ANN) are gaining prominence as a method of classification in a wide range of disciplines. In this study ANN is applied to data from a verbal autopsy study as a means of classifying cause of death.

Methods A simulated ANN was trained on a subset of verbal autopsy data, and the performance was tested on the remaining data. The performance of the ANN models were compared to two other classification methods (physician review and logistic regression) which have been tested on the same verbal autopsy data.

Results Artificial neural network models were as accurate as or better than the other techniques in estimating the cause-specific mortality fraction (CSMF). They estimated the CSMF within 10% of true value in 8 out of 16 causes of death. Their sensitivity and specificity compared favourably with that of data-derived algorithms based on logistic regression models.

Conclusions Cross-validation is crucial in preventing the over-fitting of the ANN models to the training data. Artificial neural network models are a potentially useful technique for classifying causes of death from verbal autopsies. Large training data sets are needed to improve the performance of data-derived algorithms, in particular ANN models.

KEY MESSAGES

• Artifical neural networks have potential for classifying causes of death from verbal autopsies.

• Large datasets are needed to train neural networks and for validating their performance.

• Generalizability of neural network models to various settings needs further evaluation.

Verbal autopsies, classification, neural networks

In many countries routine vital statistics are of poor quality, and often incomplete or unavailable. In countries where vital registration and routine health information systems are weak, the application of verbal autopsy (VA) in demographic surveillance systems or cross-sectional surveys has been suggested for assessing cause-specific burden of mortality. The technique involves taking an interviewer-led account of the symptoms and signs that were present preceding the death of individuals from their caretakers. Traditionally the information obtained from caretakers is analysed by physicians and a cause(s) of death is reached if a majority of physicians on a panel agreed on a cause(s). The accuracy of physician reviews has been tested in several settings using causes of death assigned from hospital records as the ‘gold standard’. Although physician reviews of VA gave robust estimates of cause-specific mortality fractions (CSMF) of several causes of death, the sensitivity, specificity and predictive values varied between causes of death and between populations1,2 and had poor repeatability of results.3

Arguments to introduce opinion-based and/or data-derived algorithm methods of assigning cause of death from VA data are based on both the quest for accuracy and consistency, as well as the logistical difficulties in getting together a panel of physicians to review what are often large numbers of records. However, physician review performed better than set diagnostic criteria (opinion-based or data-derived) given in an algorithm to assign a cause of adult death.4 One promising approach to diagnose disease status has been artificial neural networks (ANN) which apply non-linear statistics to pattern recognition. For example, ANN predicted outcomes in cancer patients better than a logistic regression model.5 Duh et al. speculate that ANN will prove useful in epidemiological problems that require pattern recognition and complex classification techniques.6 In this report, we compare the performance of ANN and logistic regression models and physician review for reaching causes of adult death from VA.

## Methods

### An overview of neural networks

Although often referred to as black boxes, neural networks can in fact easily be understood by those versed in regression analysis techniques. In essence, they are complex non-linear modelling equations. The inputs, outputs and weights in a neural network are analogous to the input variables, outcome variables and coefficients in a regression analysis. The added complexity is largely the result of a layering of ‘nodes’ which provides a far more detailed map of the decision space. A single node neural network will produce a comparable output to logistic regression, where a function will combine the weights of the inputs to produce the output (Figure 1).

Combining these nodes into multiple layers adds to the complexity of the model and hence the discriminatory power. In so doing, a number of elements, each receiving all of the inputs and producing an output, have these outputs sent as inputs to a further element(s). The architecture is called a multi-layer perceptron (Figure 2).

The study population and field procedures of the VA data used in this analysis are described elsewhere.1 In brief, data were collected at three sites (a regional hospital in Ethiopia, and two rural hospitals in Tanzania and Ghana). Adults dying at these hospitals who lived within a 60-km radius of the institution were included in the study. A VA questionnaire was administered by interviewers with at least 12 years of formal education. The reference diagnoses (gold standard) were obtained from a combination of hospital records and death certificates by one of the authors (DC) together with a local physician in each site. A panel of three physicians reviewed the VA data and reached a cause of death if any two agreed on a cause (physician review).

The method used to derive algorithms from the data using logistic regression models has been described elsewhere.4 Each subject was randomly assigned to the train dataset (n = 410) or test dataset (n = 386), such that the number of deaths due to each cause (gold standard) was the same in both datasets. If a cause of death had odd numbers, the extra subject was included in the train dataset. Symptoms (includes signs) with odds ratio (OR) ≥2 or ≤0.5 in univariate analyses were included in a logistic model and then those symptoms that were not significant statistically (P > 0.1) were dropped from the model in a backward stepwise manner. Coefficients of each symptom remaining in the model were summed to obtain a score for each subject i.e. Score = b1×1+b2×2+…, where bixi are the log OR bi of symptoms xi in the model. A cut-off score was identified for each cause of death (included 16 primary causes of adult death) that gave the estimated number of deaths closest to the true number of cause-specific deaths, such that the sensitivity was at least 50%.

We used the same train and test datasets used by Quigley et al. for training and testing an ANN. The data were ported to Microsoft Excel™ and analysed using NeuroSolutions 3.0™ (Lefebvre WC. NeuroSolution Version 3.020, Neurodimension Inc.1994. [www.nd.com]). All models were multi-layer perceptrons with a single hidden layer and trained with static backpropogation. The number of nodes in the hidden layer were varied according to the number of inputs and network performance. A learning rate of 0.7 was used throughout with the momentum learning rule. A sigmoid activation function was used for all processing elements.

Model inputs were based on those used in the logistic regression study, with further variables added to improve discrimination in instances when they improved the model performance. Sensitivity analysis provided the basis for evaluating the role of the inputs in the models.

For each diagnosis, the first 100 records of the training subset were used in the first training run of each model as a cross-validation set to determine the optimal number of hidden nodes and the training point at which the cross-validation mean squared error reached a trough. Thereafter the full training set was used to train the network to this point.

The output weights were then adjusted by a variable factor until the CSMF was as close as possible to 100% of the expected value in testing runs on the training set. At this point the network was tested on the unseen data in the test subset.

Weighted (by number of deaths) averages for sensitivity and specificity were calculated for each method. A summary measure for CSMF was calculated for each method by summing the absolute difference in observed and estimated number of cases for each cause of death, dividing by the total number of deaths, and converting to a percentage.

## Results

Table 1 shows the comparison of validity of the logistic regression models versus the ANN models for estimating CSMF by comparing estimated with observed number of cases as well as sensitivity and specificity.

The CSMF was estimated to within 10% of the true value in 8 out of 16 classes (causes of death) by the ANN. In a further six classes it was estimated to within 25% of the true value. In the remaining two classes the CSMF was extremely low (tetanus and rabies). The summary measure for CSMF favours those methods that are more accurate on the more frequently occurring classes and may mask poor performance on rare causes of death. In this measure however, calculated from the absolute number of over- or under-diagnosed cases, the neural network method performed better than logistic regression models (average error 11.27% versus 31.27%), and compared well with physician review (average error of 12.84%). In the assessment of chance agreement between ANN and gold-standard diagnoses, the kappa value was ≥0.5 for the following classes: rabies (0.86), injuries (0.76), tetanus (0.66), tuberculosis and AIDS (0.55), direct maternal causes (0.52), meningitis (0.50), and diarrhoea (0.50).

There was a trade-off between specificity and sensitivity, and in some instances the neural network performed better than other techniques in one at the expense of the other. Compared to logistic regression, the networks performed better in both parameters for tuberculosis and AIDS, meningitis, cardiovascular disorders, diarrhoea, and tetanus. They produced a lower sensitivity for malaria (compared with logistic regression), but a higher specificity. The overall and disease-specific sensitivities and specificities compared favourably with logistic regression, but did not match the performance of physician review.

## Discussion

### Accuracy of CSMF estimates

One of the most significant findings of this analysis is the relative accuracy in assessing the fraction of deaths that are due to specific causes, especially for the more frequently occurring classes. The accuracy in this estimate does not always correlate with the reliability estimated by the kappa statistic. Care was taken to find a weighting for the output that would lead to a correct CSMF in the training set. The choice of this weight is analogous to selecting the minimum total score at which a case is defined in the logistic regression models. This then led to surprisingly good estimates in the testing set. It is a feature of the train and test subsets however that the number of members in each class is similar. Manipulating either subset so that the CSMF differed, by randomly removing or adding records of the class in question, did not alter the accuracy of the CSMF estimates if the number of training examples for the class was not decreased in the training subset. With less frequently occurring classes such as pneumonia, decreasing the number of training examples in the training set, reduced the accuracy of the CSMF estimate. This is essentially an issue of generalisation, and it is to be expected that networks that are trained with fewer examples are less likely to be generalizable. It is suggested that it is for this reason that the CSMF estimates for the five most frequently occurring classes are all within 10% of the expected values. It would be expected furthermore that if the datasets were larger, that the generalizability of the CSMF estimates for the less frequently occurring classes would improve.

At the stage of data analysis the question can be asked as to whether or not there is an output level above which class membership is reasonably certain, and below which misclassification is more likely to occur. Looking at the tuberculosis-AIDS model (n = 71), as well as the meningitis model (n = 32), and ranking the top 20 test outputs in descending order by value (reflecting the certainty of the classification), 13/20 of these outputs correctly predict the class membership in both instances. The sensitivities for the models overall were 66% and 56% respectively. The implication is that without a gold standard result for comparison, it would be difficult to delineate the true positives from the false positives even in the least equivocal outputs. This is in keeping with observations that different data-derived methods arrive at their estimates differently. One study to predict an acute abdomen diagnosis from surgical admission records demonstrated that data-derived methods with similar overall performance correlated poorly as to which of the records they were correctly predicting.7

### Mechanisms of improved performance

A single layer neural network (i.e. a network with only inputs, and one processing element) is isomorphic with logistic regression. A network with no hidden nodes produced almost identical results when comparing the input weights to the log(OR) for the four inputs used in the regression model to predict malaria as the cause of death. In those instances where the performance of logistic regression and neural network models differ, it is of interest as to know the mechanisms by which improvements are made. The results from this study indicate that the differences in performance of the neural networks are achieved both by improved fitting of those variables already known to be significantly predictive of class membership, through the modelling of interaction between them, and by additional discriminating power conferred by variables that are not significantly predictive on their own

The first mechanism was borne out in one of the meningitis models in which the exact same inputs used in the logistic regression model were used in the neural network model with an improvement in performance. Exploring the sensitivity analysis for cardiovascular deaths (Table 2), the network outputs are surprisingly sensitive to the absence of a tuberculosis history, which was not strongly predictive by itself. Age above 45 years old was the seventh most predictive input in the regression model, whereas it was the input to which the neural network model was second most sensitive. In the case of meningitis, presence of continuous fever was more important in the regression model, whilst presence or absence of recent surgery and abdominal distension were more significant in the ANN model (Table 3). The network has mapped relationships between the inputs that were not predicted by the regression model.

### Effect of size of dataset

Both data-derived methods stand to benefit from more training examples. In the regression models, some inputs not currently utilized may yield significant associations with outputs when larger datasets are used. With enough nodes and training time, it was possible in the course of this analysis to train a neural network to completely map the training set with 100% sensitivity and specificity. However, this level of sensitivity and specificity was not reproduced when these models were tested in the test dataset. What it did demonstrate is the ability of the method to map complex functions. The key point is one of generalizability. In the models presented above, training was stopped and the nodes limited to ensure that the generalizability was not compromised. With more training examples, it is likely that the networks would develop a better understanding of the relationships between inputs and outputs before over-training occurs. Arguably, the neural network models would stand to improve performance more than the regression models should larger training sets be available. However, further training may not achieve algorithms of sufficiently high sensitivity and specificity to obviate the need for algorithms with particular operating characteristics suitable for use in specific environments.

### Physician review

Only 78% of the reference diagnoses were confirmed by laboratory tests. Since 22% of the reference diagnoses were based on hospital physicians' clinical judgement, it is not surprising that physician review of VA performed better than the other methods. Nevertheless, physician review remains the optimal method of analysis, as far as overall performance is concerned, for gathering cause-specific mortality data as good as the data produced by routine health information systems.1 The technique by which physicians in this study came to their classification differed considerably, as they made extensive use of the open section of the questionnaire from which information was not coded for analysis by the other techniques. Interestingly though, other methods are able to come close if the CSMF is used as the outcome of choice, as indeed it often is. Thus ANN or logistic regression models based algorithms have the potential for substituting physician review of VA.

### Limitations of the technique

At various points we have alluded to some of the difficulties and limitations of using neural networks for the analysis. These are summarized in Table 4.

Even with sensitivity analysis, we had no way of working out which were going to be the most important inputs prior to creating a model and conducting a sensitivity analysis on it. There is some correlation with linearly predictive inputs that helps in the initial stages.

Determining the weighting for the output for providing the optimum estimate of the CSMF was time-consuming. The software provides an option for prioritizing sensitivity over specificity, but no way of balancing the number of false positives and false negatives that would give an accurate CSMF estimate.

Designing the optimal network topology requires building numerous networks in search of the one with the lowest least mean squared error. The number of hidden nodes, inputs and training time all affect the performance of the network. Whilst training is relatively quick compared to the many hours it took to train ANN in the early days of their development, it is still time-consuming to build and train multiple networks for each model.

Cross-validation to prevent over-training required compromising the number of training examples to allow for a cross-validation dataset.

Sensitivity and specificity of the ANN algorithms were not high enough to be generalizable to a variety of settings. Furthermore, the accuracy of individual and summary estimates of CSMF obtained in this study could be due to the similarity in the CSMF between the training and test datasets. Thus large datasets from a variety of settings are needed to identify optimal algorithms for each site with different distributions of causes of death.

## Conclusions

Classification software based on neural network simulations is an accessible tool which can be applied to VA data potentially outperforming other the data-derived techniques already studied for this purpose. As with other data-derived techniques, over-fitting to the training data leading to a compromise in the generalizability of the models is a potential limitation of ANN. Increasing the number of training examples is likely to improve performance of neural networks for VA. However, ANN algorithms with particular operating characteristics would be site-specific. Thus optimal algorithms need to be identified for use in a variety of settings.

Table 1

Comparison of performance of physician review, logistic regression and neural network models

Combined datasetTest dataset
Physician review Logistic regression Neural network
Causes of deathObserved Est.aSens.bSpec.cObserved Est.aSens.bSpec.cEst.aSens.bSpec.cKappa
TB-AIDSd148 15176 94 71 7865 90 7766 90 0.55
Malaria 85 8033 93 42 89 48 80 4429 91 0.19
Meningitis 66 6059 97 32 3353 95 3456 95 0.50
CVSe disorders 65 5748 96 32 3234 94 3047 97 0.44
Acute abdominal conditions 55 6569 97 27 18 37 98 2948 96 0.42
Diarrhoeal diseases 51 5861 96 25 3048 95 3160 96 0.50
Direct maternal causes 50 5882 98 25 18 52 97 1948 98 0.52
Neoplasms 34 3650 98 16 24 19 94 1496 0.03
Injuries 33 3697 99.5 15 1680 99 1680 99 0.76
Hepatitis 32 14 34 99.6 16 10 97 1995 0.01
Chronic liver disease 25 2940 98 12 1497 1525 97 0.19
Anaemia 24 12 33 99.5 12 25 25 94 1117 98 0.15
Pneumonia 23 1939 98.7 11 14 27 97 1327 97 0.23
Renal disorders 21 2238 98 10 1010 98 1010 98 0.08
Tetanus 13 1077 100 517 99 50 100 0.66
Rabies 786 99.9 100 99.7 100 0.86
Weighted averagef58.95 96.39 43.98 93.01 45.32 94.72
Summary CSMFg accuracyh12.84 31.27 11.83
Combined datasetTest dataset
Physician review Logistic regression Neural network
Causes of deathObserved Est.aSens.bSpec.cObserved Est.aSens.bSpec.cEst.aSens.bSpec.cKappa
TB-AIDSd148 15176 94 71 7865 90 7766 90 0.55
Malaria 85 8033 93 42 89 48 80 4429 91 0.19
Meningitis 66 6059 97 32 3353 95 3456 95 0.50
CVSe disorders 65 5748 96 32 3234 94 3047 97 0.44
Acute abdominal conditions 55 6569 97 27 18 37 98 2948 96 0.42
Diarrhoeal diseases 51 5861 96 25 3048 95 3160 96 0.50
Direct maternal causes 50 5882 98 25 18 52 97 1948 98 0.52
Neoplasms 34 3650 98 16 24 19 94 1496 0.03
Injuries 33 3697 99.5 15 1680 99 1680 99 0.76
Hepatitis 32 14 34 99.6 16 10 97 1995 0.01
Chronic liver disease 25 2940 98 12 1497 1525 97 0.19
Anaemia 24 12 33 99.5 12 25 25 94 1117 98 0.15
Pneumonia 23 1939 98.7 11 14 27 97 1327 97 0.23
Renal disorders 21 2238 98 10 1010 98 1010 98 0.08
Tetanus 13 1077 100 517 99 50 100 0.66
Rabies 786 99.9 100 99.7 100 0.86
Weighted averagef58.95 96.39 43.98 93.01 45.32 94.72
Summary CSMFg accuracyh12.84 31.27 11.83

View Large

Table 2

Comparison of the most important inputs for two data-derived models for assigning cardiovascular deaths

Logistic regression modelNeural network model
RankaInput RankbInput
Puffiness of face Puffiness of face
Cough 3–14 days Age ≥45years
Abdominal distension 8–30 days No weight loss
No weight loss Abdominal distension 8–30 days
No jaundice History of hypertension
History of hypertension No history of tuberculosis
Age ≥45years No jaundice
No recent surgery
Pallor
10 No stiff neck
11 Cough >3 weeks
12 Shortness of breath
13 No chronic diarrhoea
14 No productive cough
15 Chest pain
16 Wheeze
17 No Continuous fever
Logistic regression modelNeural network model
RankaInput RankbInput
Puffiness of face Puffiness of face
Cough 3–14 days Age ≥45years
Abdominal distension 8–30 days No weight loss
No weight loss Abdominal distension 8–30 days
No jaundice History of hypertension
History of hypertension No history of tuberculosis
Age ≥45years No jaundice
No recent surgery
Pallor
10 No stiff neck
11 Cough >3 weeks
12 Shortness of breath
13 No chronic diarrhoea
14 No productive cough
15 Chest pain
16 Wheeze
17 No Continuous fever

View Large

Table 3

Comparison of the most important inputs for two data-derived models for assigning death due to meningitis

Logistic regression modelNeural network model
RankaInput RankbInput
Stiff neck Stiff neck
No cough Cough
No pallor Recent surgery
Continuous fever Pallor
Stiff body Abdominal distension
Injury/accident
Severe loss of weight
Body stiffness
Tuberculosis
Logistic regression modelNeural network model
RankaInput RankbInput
Stiff neck Stiff neck
No cough Cough
No pallor Recent surgery
Continuous fever Pallor
Stiff body Abdominal distension
Injury/accident
Severe loss of weight
Body stiffness
Tuberculosis

View Large

Table 4

Limitations of the artificial neural network technique

 Selecting inputs is not straightforward Prioritizing cause-specific mortality fraction over sensitivity or specificity is a manual process Designing optimal networks for each cause of death is time-consuming Sensitivity and specificity may be not high enough for the algorithms to be generalizable to a variety of settings
 Selecting inputs is not straightforward Prioritizing cause-specific mortality fraction over sensitivity or specificity is a manual process Designing optimal networks for each cause of death is time-consuming Sensitivity and specificity may be not high enough for the algorithms to be generalizable to a variety of settings

View Large

## References

1

Chandramohan D, Maude H, Rodrigues L, Hayes R. Verbal autopsies for adult deaths: their development and validation in a multicentre study.

Trop Med Int Health

1998

;

3

:

436

–46.

2

Snow RW, Armstrong ARM, Forster D et al. Childhood deaths in Africa: uses and limitations of verbal autopsies.

Lancet

1992

;

340

:

351

–55.

3

Todd JE, De Francisco A, O'Dempsey TJD, Greenwood BM. The limitations of verbal autopsy in a malaria-endemic region.

Ann Trop Paediatr

1994

;

14

:

31

–36.

4

Quigley M, Chandramohan D, Rodrigues L. Diagnostic accuracy of physician review, expert algorithms and data-derived algorithms in adult verbal autopsies.

Int J Epidemiol

1999

;

28

:

1081

–87.

5

Jefferson MF, Pendleton N, Lucas SB, Horan MA. Comparison of a genetic algorithm neural network with logistic regression for predicting outcome after surgery for patients with nonsmall cell lung carcinoma.

Cancer

1997

;

79

:

1338

–42.

6

Duh MS, Walker AM, Pagano M, Kronlund K. Epidemiological interpretation of artificial neural networks.

Am J Epidemiol

1998

;

147

:

1112

–22.