Que1.12. What are the advantages and disadvantages of different types of machine literacy algorithms? Answer Advantages of a supervised machine learning algorithm 1. Classes represent the features on the ground. 2. Training data is applicable unless features change. Disadvantages of a supervised machine learning algorithm 1. Classes may not match spectral classes. 2. Varying thickness in classes. 3. Cost and time are involved in opting training data. Advantages of an unsupervised machine learning algorithm 1. No former knowledge of the image area is needed. 2. The occasion for mortal error is minimized. 3. It produces unique spectral classes. 4. fairly easy and fast to carry out. Disadvantages of an unsupervised machine learning algorithm 1. The spectral classes don’t inescapably represent the features on the ground. 2. It doesn’t consider spatial connections in the data. 3. It can take time to interpret the spectral classes. Advantages of asemi-supervised machine learning algorithm 1. It’s easy to understand. 2. It reduces the quantum of annotated data used. 3. It’s stable, fast convergent. 4. It’s simple. 5. It has high effectiveness. Disadvantages ofsemi-supervised machine literacy algorithm 1. replication results aren’t stable. 2. It isn’t applicable to network position data. 3. It has low delicacy. Advantages of underpinning learning algorithm 1. underpinning literacy is used to break complex problems that can not be answered by conventional ways. 2. This fashion is preferred to achieve long- term results which are veritably delicate to achieve. 3. This literacy model is veritably analogous to the literacy of mortal beings. Hence, it’s close to achieving perfection. Disadvantages of underpinning learning algorithm 1. Too important underpinning literacy can lead to an load of countries which can dwindle the results. 2. underpinning literacy isn’t preferable for working simple problems. 3. underpinning literacy needs a lot of data and a lot of calculation. 4. The curse of dimensionality limits underpinning learning for real physical systems. Que1.13. Write short note on Artificial Neural Network( ANN). Answer 1. Artificial Neural Networks( ANN) or neural networks are computational algorithms that intended to pretend the geste
of natural systems composed of neurons. 2. ANNs are computational models inspired by an beast’s central nervous systems. 3. It’s able of machine literacy as well as pattern recognition. 4. A neural network is an oriented graph. It consists of bumps which in the natural analogy represent neurons, connected by bends. 5. It corresponds to dendrites and synapses. Each bow associated with a weight at each knot. 6. A neural network is a machine learning algorithm grounded on the model of a mortal neuron. The mortal brain consists of millions of neurons. 7. It sends and process signals in the form of electrical and chemical signals. 8. These neurons are connected with a special structure known as synapses. Synapses allow neurons to pass signals. 9. An Artificial Neural Network is an information processing fashion. It works like the way mortal brain processes information. 10. ANN includes a large number of connected processing units that work together to reuse information. They also induce meaningful results from it. Que1.14. Write short note on clustering. Answer 1. Clustering is a division of data into groups of analogous objects. 2. Each group or cluster consists of objects that are analogous among themselves and different to objects of other groups as shown inFig.1.14.1. 1.14.1. Clusters. 3. A cluster is a collection of data objects that are analogous to one another within the same cluster and are different to the object in the other cluster. 4. Clusters may be described as connected regions of a multidimensional space containing fairly high viscosity points, separated from each other by a region containing a fairly low viscosity points. 5. From the machine learning perspective, clustering can be viewed as unsupervised literacy of generalities. 6. Clustering analyzes data objects without help of known class marker. preface 1 – 16 L( CS/ IT- Sem- 5) 7. In clustering, the class markers aren’t present in training data simply because they aren’t known to cluster the data objects. 8. Hence, it’s the type of unsupervised literacy. 9. For this reason, clustering is a form of literacy by observation rather than learning by exemplifications. 10. There are certain situations where clustering is useful. These include a. The collection and bracket of training data can be expensive and time consuming. thus it’s delicate to collect a training data set. A large number of training samples aren’t all labelled. also it is useful to train a supervised classifier with a small portion of training data and also use clustering procedures to tune the classifier grounded on the large, unclassified dataset. b. For data mining, it can be useful to search for grouping among the data and also fete the cluster. c. The parcels of point vectors can change over time. also, supervised bracket isn’t reasonable. Because the test point vectors may have fully different parcels. d. The clustering can be useful when it’s needed to search for good parametric families for the class tentative consistence, in case of supervised bracket. Que1.15. What are the operations of clustering? Answer Following are the operations of clustering 1. Data reduction a. In numerous cases, the quantum of available data is veritably large and its processing becomes complicated. Cluster analysis can be used to group the data into a number of clusters and also reuse each cluster as a single reality. c. In this way, data contraction is achieved. 2. thesis generation a. In this case, cluster analysis is applied to a data set to infer thesis that enterprises about the nature of the data. Clustering is used then to suggest thesis that must be vindicated using other data sets. 3. thesis testing In this environment, cluster analysis is used for the verification of the validity of a specific thesis. 4. vaticination grounded on groups a. In this case, cluster analysis is applied to the available data set and also the performing clusters are characterized grounded on the characteristics of the patterns by which they’re formed. In bracket, data are grouped by assaying the data objects whose class marker is known. There’s some previous knowledge of the attributes of each bracket. It’s done by classifying affair grounded on the values of the input data. The number of classes is known before bracket as there is predefined affair grounded on input data. It’s considered as the supervised literacy because class markers are known before. 1. Clustering analyzes data objects without known class marker. 2. There’s no previous knowledge of the attributes of the data to form clusters. 3. It’s done by grouping only the input data because affair is not predefined. 4. The number of clusters is not known before clustering. These are linked after the completion of clustering. 6. It’s considered as unsupervised literacy because there’s no previous knowledge of the class markers. b. In this sequence, if an unknown pattern is given, we can determine the cluster to which it’s more likely to belong and characterize it grounded on the characterization of the separate cluster.
