Scribd
Upload
66 views57 pages

ML Unit 5

Aktu MAchine Learning Quantum

Uploaded by

Yash Vardhan Singh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
66 views57 pages

ML Unit 5

Aktu MAchine Learning Quantum

Uploaded by

Yash Vardhan Singh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 57
Part-1 Part-2 : Part-3 Part-4 Part-5 Part-6 Reinforce, Learning ent Genetic Algority.” h Introduction to Reinforcement or 5-2P Learning to 5.5p Learning Task -- Example of Reinforcement .---- 5-7P ti Learning ip Practice ‘0 5-8P sion of Reinforcement 5-13P to 5-13P Introduction to Deep Q- Learning ---- yithm. Introd 5-15P to 5-19P Genetic Algo ccs S-11P to 5-18P © a KNN | ; 7. | Applications Robotics, Machine | Risk evaluation, foree; | . 7 : » forecast, | | Learning, Aireraft|sales, object recognition, | L control, AL " What is reinforcement learning ? Explain passive Que 5.3. reinforcement learning and active reinforcement learning. Answer ] Reinforcement learning : Refer Q. 5.1, Page 5-2P, Unit-5. Passive reinforcement learning : Passive reinforcement learning occurs when the agent does not have control over its actions. Instead, the actions are determined by an external agent, such as a human operator or a pre-programmed algorithm. enalties based 2 Inthese situations, the agent simply receives rewards or p' . on its actions, and its goal is to learn from this feedback to improve its performance. 3. Agood example of passive reinforcement learning is in robotics, where an external agent might provide rewards for reaching a taret location L 4) Reint .Bom-4) Oreoman, poe a Wnt A... cific task Seite stg Jeting # spect 7 The tg ro com? ‘a reward or penalty top ite 7 i, form to choose its own actions, jd pet? e | Acti | } gctio! m toy, pte ynforcement learning; have the inforcement learning is wh ve rein en the ra to perform based on the eisrrany Staten i ' Y chose, eans that the agent has com, st, OF the ony: i 8 the | i i ‘ rte ment, ely: mt explore different options tg teeming "Satan and | . the be is 4g roWal" AY to max i ; timiza for example, ina oe ACtIVE tein | 3 (yould be reaponsl le . leciding whic 0 = tearsng agent | valuation of the current state of the game...” Basey nie Fig. 5.17.1. 4, ona higher level, Deep Q learning works as such : Gather and store samples in a replay buffer with current policy j. Random sample batches of experiences from the replay butter. ii, Use the sampled experiences to update the @ network. i iv. Repeat 1-3. Que 5.18. | What are the steps involved in deep Q-learning network Answer using deep Q-learning networks : Steps involved in reinforcement learning 1. Allthe past experience is stored by the user in memory. 2. The next action is determined by the maximum outps! ofthe Q-network. 9 Machine Learning SSE Oty of the © Predicteg oe) The loss function here is mean squared error o: and the target Q-value ~ Q*. This is basically a regress; ion p 4. However, we do not know the target or actual vale © he oben dealing with a reinforcement learning problen; Goin ere a, ne pa back . Ary Q-value update equation derived from the Bellman ¢ ) A,) max 7 the MS, A) ASA) + ALR. + MAAS, 0) QS Ay rhage Que 5.19. ] Write pseudocode for deep Q. learning, nea Start with Q,(s, a) for all s, a. Get initial state s For k = 1, 2, ... till convergence Sample action a, get next state s' Ifs' is terminal : target = R(s, Bs Sample new initial state s’ target = R(s,a,s')+7 raxQ,(e/, a’) 8,21 — 0, ~ AVE piers al @a(s, a)- re ee Genetic Algorithm, Introduction. — ' Que 5.20. | Write short note on Genetic algorithm. Refer Q. 1.32, Page 1-30P, Unit-1. Que 5.21. ] Write procedure of Genetic algorithm with advantages and disadvantages. Answer ] Procedure of Genetic algorithm : 1 Generate a set of individuals as the initial population 2. Use genetic operators such as selection or cross over else ses’ Reinforcement por ne eg " ati + Algrihn ‘i a yo ion oF digital reverse aegar, jo got? ene fonction of the new VY ol ene PODUMAtinn ate ga fur stem sia et fine ‘embers from the original Popa dividuals and ceplace (jefe) g. and terminate when inl Oredefined 20 Mlation met. enetic algorithm : et be executed 198° ocithmscan 'n parallel. Ha o say BeMetic algorithm or solving optimization problems, of Genetic algorithm : 6 4. 5 i ‘on of the fitness function is difficaty pea" 48 it depends on the 7 ae "jon of guitable genetic operators i diffu, flowchart of GA 2 “ 2. praw 4 and explain the Working (Cu ‘ rithm : Rete Q 1,32, Page 1-30P, Unit-1, 8" ne principle! i the “ne principle of GA, we Consider unconstrained 4 ie elf to minimize f(X), for f(X) > 0, then we can write the objective "function as * eee maxim? 14 FQ) 4, AX <0 instead of minimizing AX), maximize |-/(X)|. Hence, both maximization and minimization problems can be handled by GA. Que 5.23. | Write short notes on procedures of GA. Answer 1. Start : Generate random population of n chromosomes. 2 Fitness ; Evaluate the fitness fix) of each chromosome ¢ in the population. Machine Learning 3 New pop m : Create a new popul steps until the new population is complete : a, Selection : Select two parent chromosomes from a Populati, according to their fitness, * b. Crossover : With a crossover probability crossover the Parent, to form new offspring (children). If no crossover was Performed offspring is the exact copy of parents, 7 ce Mutation : With a mutation probability mutate new offspring at each locus (position in chromosome). ds Accepting : Place new offspring in the new Population. 4. Replace : Use new ace generated population for a further run of the algorithm. 5. Test: Ifthe end condition is satisfied, stop, and return the best Solution in current population. 6. Go to step 2. Que 5.24. What are the benefits of using GA ? What are its limitations ? Answer ] Benefits of using GA: 1. Itis easy to understand. 2. Itis modular and Separate from application. 3. It supports multi-objective optimization, 4. Itis good for noisy environment. Limitations of genetic algorithm are: 1. The problem of identifying fitness function. Definition of representation for the problem. Premature convergence occurs. The problem of choos: population, mutation strength Cannot use gradients. Cannot easily incorporate problem specific information. Not good at identifying local optima. No effective terminator. Aw LV ing the various parameters like the size of the rate, crossover rate, the selection method and its Sr on Reinfore: y f ement Learning & Genetic Algorithm wea oC asem-4) ffecti¥® for smooth unimodal functions. oo 7 ‘ °. ae to be coupled with a local search technique. Nee 10. 5.25- write short notes of genetic representations. we aa ewer ve An neti€ representation 18 a way of representing solutions/individuals 1, 7 glutionary ‘computation methods. iduals in ev , epresentation can encode appearance, behavior. physical . physica etic repre’ 2 Gest of individuals All the individuals of a population are represented by using binary 3. encoding. permutational encoding, encoding bytree. binary netic algorithms usé linear binary representations: The most staniard 4 method of representation is an array of bits. These genetic representations are convenient because parts of individual 6 are easily aligned due to their fixed size which makes simple veer operation. Que 5.26. Give the detail of genetic representation (Encoding). OR Explain different types of encoding in genetic algorithm. Answer Genetic representations : 1, Encoding: ‘a’. Encoding isa process of representing individual genes. b.. The process can be performed using bits, numbers, trees, arrays lists or any other objects. 2 ¢. The encoding depends mainly on solving the problem Binary encoding : * Se mmenly used method of genetic a Binary encoding is the most representation because GA uses b. Inbinary encoding, every chromosome is # Chromosome A|10110010110010101 110010) Chromosome B/11111110000011000001111? Bi comes inary encoding gives many possible chromosome’ Lo a etal or Hexadecimal encoding‘ "The encoding is done using octal or hexadecimal aumbers this type of encoding. string of bits. or 1. Machine Learning 5-19 P (MCA-Sen, i) sina ba Chromosome Octal Hexadecimal Chromosome A 54545345 _ B2CAE5 Chromosome B | 77406087 | FEOCIF — 4 Permutation encoding (real number encoding) : a. Permutation encoding can be used in ordering problems, su, Travelling Salesman Problem (TSP). b. In permutation encoding, every chromosome is a string of numbers which represents number in a sequence. : | Chromosome A/15 3264798 Chromosome B|856723149 5. Value encoding: a. Direct value encoding can be used in problems, where some complicated values, such as real numbers, are used. In value encoding, every chromosome is a string of some values, c. Values can be anything connected to problem, real numbers or chars to some complicated objects. Chromosome A/1.2324 5.3243 0.4556 2.3293 2.4545 Chromosome B|ABDJEIFJDHDIERJFDLDFLFEGT ch as Chromosome C/(back), (back), (right), (forward), (left) 6& Tree encoding: a Tree encoding is used for evolving programs or expressions, for genetic programming. b. In tree encoding, every chromosome is a tree of some objects, such as functions or commands in programming language. c. Programming language LISP is often used to this, because programs in it are represented in this form and can be easily parsed as a tree, so the cross-over and mutation can be done relatively easily. Chromosome A Chromosome B | (2) Do_until | SD yy ® Step Wall (+xV/5y) ji eee sat MS mnt Learning & Gertie Algorithm EN oa 5 Pa Components. Za | Draw genetics cycle for genetic algorithm, : que 5 OR ne rototypical genetic algorithm alon ert ible in it. pesratio ns possi oR enetic algorithm cycle of reproduction with its % with various in 8 explain comsponsent®- ARTU 2021-22, Marks 10] Answer : Generational cycle of GA: Population (Chromosomes) Decoded Offsprings string New generation Genetic Byaluation operator i (Fitness) Manipulation Mate Reproduction Selection (Mating pool) Fig. 5.27.1. The GA cycle. Components of generational cycle in GA: 1 Population (Chromosomes) : A population is collection of individuals. A population consists of a number of individuals being tested, the phenotype parameters defining the individuals and some informanon about search space. | Evaluation (Fitness) : A fitness function is a particular type of objective function that quantifies the optimality of a solution (i. 8 Sr im a genetic algorithm so that particular chromosome may ber ®gainst all the other chromosomes. Selection : During each successive generation, cee Population is selected to breed a new gen ‘utions are selected through a fitness-based process. a proportion of the eration. Individual ha eS eis. 4. Generic operator : A genetic operator is an operator used in genotic algorithm to guide the algorithm towards a solution to agiven problem Part-10 GA Cycle of Reproduction. Que 5.28. | How to identify the reproduction cycle of genetic algorithm ? Explain with suitable example. Answer Reproduction cycle of genetic algorithm : 1. The reproduction cycle of a genetic algorithm refers to the process of creating a new generation of candidate solutions from the previous generation using genetic operators. 2. The reproduction cycle typically consists of the following steps : i. Selection : Select a subset of the fittest individuals from the previous generation to become parents for the next generation, ii Crossover : Create new individuals by combining the genetic material of the selected parents through a crossover operation. iii. Mutation : Introduce random changes into the genetic material of the new individuals through a mutation operation. iv. Replacement: Replace the least fit individuals from the previous generation with the new individuals to form the next generation. 3. The reproduction cycle is repeated for a fixed number of generations or until a termination criterion is met, such as reaching a maximum fitness or a maximum number of generations. Example: 1. Consider a genetic algorithm for optimizing the parameters ofa machine learning model. 2. The genetic algorithm starts with a randomly initialized population of candidate solutions, each of which corresponds to a different set of parameters for the model. 3. Ineach generation, the fitness of each individual is evaluated. _ The fittest individuals are selected as parents for the next gener 5, The new individuals are created by combining their genetic material through crossover and introducing random changes through mutation. 6. Thenew individuals are then evaluated for fitness and used to form the next generation. ation. eee me a-Sem-4) Reinforcement Lenrnin Pp " 8 & Genet ts 1 Alotiritive, f rocess is repeated for a fixed n, ee . ; Lumber of _. 1 hie in ‘ation criterion is met. BeNerations oF inti , tel A i ‘oducti this reproduction cycle, the Reneti following _ Benetic algorithm ; 1. BY ro" ‘e the performance of the machine learning mo, dei Bradually Se Crossover, Mutation, Ce —, que 6.29. | Explain different phases of genetic algorithm, aa | pifferent phases of genetic algorithm are: i 1 a. 2 Nae FA (Factor Analysis) fitness function : a. Initial population : The process begins with a set of individuais population. Each individual is a solution to the problem we want to solve An individual is characterized by a set of Parameterg 'varnables known as genes. Genes are joined into a string to form a chromosome ‘soluti Ina genetic algorithm, the set of genes of an individual 's nepresente:! using a string. which is called 4 Usually, binary values are used (string of 1s and 08) | Gene A383 }/1/0}]1 } } | AZT} af afi} tf] a4] | chromosome | | | 44 }ilaifolilalo | Poplin The fitness function determines how fit an individual is she ability of all individual to compete with other individual Itgives a fitness score to each individual a The pro bability that an individual will be selected for repro 's based on its fitness score. ction RE Machine Learning 6-23 P (MCA. -Bemn-4) 3% Selection: a. The idea of selection phase is to select the fittest individuals and let them pass their genes to the next generation. b. Two pairs of individuals (parents) are selected based on their fitness scores. c. Individuals with high fitness have more chance to be selected for reproduction. 4 Crossover : a. Crossover is the most significant phase in a genetic algorithm, b. For each pair of parents to be mated, a crossover point.is chosen at random from within the genes. c. For example, consider the crossover point to be 3 as shown : Al |0/0]0f;0)0/0 ‘Crossover point d= Offspring are created by exchanging the genes of parents among themselves until the crossover point is reached. A2 e. The new offspring are added to the population. A5 |}1/1/1/0/0/0 AB |0/0/0}1)1)1 5. Mutation : a. When new offspring formed, some of their genes can be subjected to a mutation with a low random probability. b. This implies that some of the bits in the bit string can be flipped. Before mutation as [ifififololo After mutation T AS 1) 1/0} 1)1)0 a , Reinforcement Learn: momsen ‘arning & Genetic Algorithm occurs to maintain jon versity wy f Muerte premature convergence Y within the population and * e" ination * ; erm ithm termin - 6 x e algorit an poring which re ‘aunt ri has converged (does a ‘odui ‘ificant}) 10 ue generation). Y different from the pret ig said that the genetic algori n it is sal gorithm has prov ations to our problem. Provided a set of —<550-| Why mutation is done in genetic algorithm ? Explain que js of mutation. types Pion is done in genetic algorithm because ; fu! . . 5 Mi It maintains genetic diversity from one generation of a population of genetic algorithm chromosomes tothe next, 3, GAcangive better solution of the problem by using mutation Types of mutation : Sy 1 Bit string mutation : The mutation of bit strings occurs through bit flips at random posititinms, Example: 1010010, % 1010110 The probability ofa mutation of a bit is 1/1,where is the length of the binary vector. Thus, a mutation rate of 1 per mutation and individual selected for mutation is reached. 2 Flip bit: This mutation operator takes the chosen genome and inverts the bits (i.e., if the genome bit is 1, it is changed to 0 and vice versa). &% Boundary : This mutation operator replaces the genome with either lower or upper bound randomly. This can be used for integer and float genes, Non-uniform : The probability that amount of mutation will go to 0 with the next generation is increased by using non-uniform mutation operator. It keeps the population from stagnating in the early stages of the evolution, Uniform : This operator replaces the value of the chosen gene with a waiform random value selected between the user-specified upper and wer bounds for that gene. ‘aussian ; This o; it Gaussian distributed random : perator adds a unit Gal ma Value to the chosen gene. Ifit falls outside of the user-specified lower or Per bounds for that gene, the new gene value is clipped an OSS :'‘93@3 Machine Learning 5-25 P (MCA, ey Sem.4) Que 5.7,] What is the main function of crossover operat Senetic algorithm ? Answer 1 ion in Crossover is the basic operator of genetic algorithm. Perform genetic algorithm depends on crossover operator. 2. Type of crossover operator used for a problem depends on the type of encoding used. 7 ance of 3. The basic principle of crossover process is to exchange genetic material of two parents beyond the crossover points. Function of crossover operation/operator in genetic algorithm ; 1. The main function of crossover operator is to introduce diversity in the Population. 2. Specific crossover made for a specific problem can improve performance of the genetic algorithm. 3. Crossover combines parental solutions to form offspring with a hope to produce better solutions. 4. Crossover operators are critical in ensuring good mixing of building blocks. 5. Crossover is used to maintain balance between exploitation and exploration. The exploitation and exploration techniques are responsible for the performance of genetic algorithms. Exploitation means to use the already existing information to find out the better solution and exploration is to investigate new and unknown solution in exploration space. PART-12 Genetic Programming, Models of Evolution and Learning. Que 5.32. | Explain different methods of selection in genetic algorithm in order to select a population for next generation. = The various methods of selecting chromosomes for parents to cross over are a. Roulette-wheel selection : Roulette-wheel selection is the proportionate reproductive method where a string is selected from the mating pool with a probability proportional to the fitness / Thus, i'” string in the population is selected with a probability proportional to F, where F is the fitness value for that string. eee i m. b Bolt i. i. i bh A) Reinforcement Learning & ¢, ‘netic Algorithm i ally kept fi _ - ae the ulation si: usu: “pt fix C . since aD of the probabilities of each string being wi Algorithm mating ol must be one. Selected for the ma The probability of the ith selected string is =a DF J a re ‘n’ is the population size. rage fitness is F=LE/Mn rl ann selection : Pi whe The ave 8 Boltzmann selection uses the concept of simulated annealing gimulated annealing is a method of functional minimization or maximization. This method simulates the process of slow cooling of molten metal to achieve the minimum function value ina minimization problem The cooling phenomenon is simulated by controlling a temperature so that asystem in thermal equilibrium at a temperature T has its energy distributed probabilistically according to P(E) = exp( - Al where ‘k’ is Boltzmann constant. This expression suggests that a systemmat a high temperature has almost uniform probability of being at any energy state, but at a low temperature it has a small probability of being at a high energy state. Therefore, by controlling the temperature T and assuming search process follows Boltzmann probability distribution, the convergence of the algorithm is controlled. Tournament selection : GA uses a strategy to select the individuals from population and insert them into a mating pool. Aselection strategy in GA is a process that favours the selection of better individuals in the population for the mating pool. There are two important issues in the evolution process of genetic Search, 1. Population diversity : Population diversity means a the Benes from the already discovered good individuals are exploited. 2. Selective pressure : Selective pressure !S the degree to which the better individuals are favoured. EO Y™SSCS”C lll Machine Learning 5-27P (MCA-Sem-4) The higher the selective pressure the better individuals are favoured. d= Rank selection : i Rank selection first ranks the population and takes every chromosome, receives fitness from the ranking. ii The worst will have fitness 1, the next 2, ..., and the best will have fitness N (N is the number of chromosomes in the population), wii The method can lead to slow convergence because the best chromosome does not differ so much from the other. e Steady-state selection : i‘ The main idea of the selection is that bigger part of chromosome should survive to next generation. ii GA works in the following way : 1. In every generation a few chromosomes are selected for creating new off springs. Que 5.33. 2 placed in that place. Then, some chromosomes are removed and new offspring is 3. The rest of population survives a new generation. Differentiate between Roulette-wheel based on fitness and Roulette-wheel based on rank with suitable example. Answer Difference: “Tt S.No. Roulette-wheel Roulette-wheel based on fitness based on rank L Population is selected with a | Probability of a population being | probability that is direetly | selected is based on its fitness proportional to their fitness | rank. | values. - _ _| 2. It computes selection | It first sort individuals in the probabilities according to their fitness values but do population according to their | fitness and then computes not preserved. not sort the individual inthe | selection probabilities according population. to their ranks rather than fitness values. ra It gives a s chance to all the | It selects the individuals with | Individuals in the population | highest rank in the population, | je be selected , Sn — _ - D4 | Diversity in the population Diversity i in the population is | | is preserved, wae auca-Sem-4) : Learning & Genetic Algorithm Je : -wheel where all px90P ne a Roulette-wheel where all chromosomes in th mag h chromosome has it © Population » ime placed each cl ite place accordingly to its fitness nction ful Chromosomes 4 Chromosomes 3: Chromosomes 1 es, Chromosomes 2 Fig. 5.33.1. Roulette-wheel selection. 9, When the wheel is spun, the wheel will finally step and pointer attached "to it will points to the one of chromosomes with bigger fitness value. 3, The different between roulette-wheel selection based on fitness and * vank is shown in Fig. 5.33.1 and Fig. 5.33.3 Chromosomes 3 ‘Chromosomes 4 Chromosomes 2 Chromosomes i Fig. 5.33.2. Situation before ranking (graph of fitnesses). Chromosomes 4 Chromosomes 3 Chromosomes 1 Chromosomes 2 Fig. 5.38.3, Situation after Fanking (graph of order numbers). = weraT] Discuss the different applications of genetic algorithms. OR ~~ Machine ‘Learning 5-29 a fg mh, Explain detail about genetic algorithm, their components a g's 8nd ¢] applications. AKTU 2092.93, iy, ay heir 810 mew] ks ¥ Genetic algorithm and components : Refer Q. 5.27, Page 5-22P Up: Application of GA: "Units, 1, Optimization : Genetic Algorithms are most co: optimization problems wherein we have to Maximize or minim: given objective function value under a given set of constraints, nize a Economics : GAs are also used to characterize variou: . Ss Econom) like the cobweb model, game theory equilibrium resolution, ase, noes ete. Ping, mmonly used j 3. Neural networks : GAs are also used to train neural netwo; particularly recurrent neural networks. rks, 4 Parallelization : GAs also have very good parallel cap: prove to be very effective means in solving certain prol provide a good area for research. 5. Image processing : GAs are used for various digital image Processing (DIP) tasks as well like dense pixel matching. 6. Machine learning : Genetics based machine learning (GBML) is a nice area in machine learning. 7. Robot trajectory generation : GAs have been used to plan the path which a robot arm takes by moving from one point to another. Que 5.35. | Explain optimization of travelling salesman problem using genetic algorithm and give a suitable example too. Answer | 1. The TSP consist a number of cities, where each pair of cities has a corresponding distance. 2. The aim is to visit all the cities such that the total distance travelled will be minimized. 3. Asolution, and therefore a chromosome which represents that solution to the TSP, can be given as an order, that is, a path, of the cities. ; 4. The procedure for solving TSP can be viewed as a process flow given in Fig. 5.35.1. The GA process starts by supplying important information such as ey of the city, maximum number of generations, population size, probabil of crossover and probability of mutation. abilities, and blems, and also on : 5 the 6. An initial random population of chromosomes is generated and fitness of each chromosome is evaluated. ——————— enetic Algorithm i Ay Reinforcement Lenrning & 6 is then transformed into a new Pr jon J ior ng three genetic operators: so, ylation Population (the next tion) using tion, crossover and jon. is used to chi tor is used to choose two ion oper : Parents from th a one een order to create a new child by crossover anne eal ti 3 i i 1 Be generation oe Proportion of the characteristics va by the good mem a Previous generation and in this a sed racteristen ate SPre8d Or he population an mize we 8 mood characteristics. ol Set GA parameters Generate initial random | population Evaluate fitness of each chromosome in the =<+———_—_. population Are optimization termination criteria met ? =, | New population L.—____| Best chromosome Parents selection for next generation Crossover of, parents chromosome Mutation of chromosome at Fig. 5.85.1, Genetic algorithm procedure for TSP. ° pe Neco generation, a new set of chromosomes where the size is L Thi ‘o the initial population size is evolved. wml nsformation process from one generation to the next continues Sceurs Wee ulation converges to the optimal solution, which usualy en a certain percentage of the population (for example 90 % > EEE ST Machine Learning 5-31P (MCA-Sem aay has the same optimal chromosome in which the best individual i as the optimal solution. Mavis taken Que 5.36, | Write short notes on convergence of genetic algorithm Answer L Agenetic algorithm is usually said to converge when there isno significant improvement in the values of fitness of the population from on generation to the next. © 2. One criterion for convergence may be such that when‘a fixed Percentage of columns and rows in population matrix becomes the same, it can be assumed that convergence is attained. The fixed percentage may be 80% or 85%. 3. In genetic algorithms as we proceed with more generations, there may not be much improvement in the population fitness and the best individual may not change for subsequent populations. 4. As the generation progresses, the population gets filled with more fit individuals with only slight deviation from the fitness of best individuals so far found, and the average fitness comes very close to the fitness of the best individuals. 5. The convergence criteria can be explained from schema point of view. 6. Aschema is a similarity template describing a subset of strings with similarities at certain positions. A schema represents a subset of all possible strings that have the same bits at certain string positions. 7. Since schema represents a robust of strings, we can associate a fitness value with a schema, i.e., the average fitness of the schema. 8. Onecan visualize’GA’s search for the optimal strings as a simultaneous competition among schema increases the number of their instances in the population. @O®O . What are the diff 89-1 P (MCA.Sem-4) i Introduction (2 Marks Questions) machine learning. Define OR Define the term learning in machine learning. AKTU 2021-22, Marks 02] AKTU 2022-23, Marks 02 hine learning is an applicatiom of artificial intelli that Mevides systems the ability to automat qineial F thigence hat from experience without being explicitly pr _ erent types of machine Yearning algorithm ? . Different types of machine learning algorithm are: Supervised machine learning algorithm ; Unsupervised machine learning algorithm . Semi-supervised machine learning algorithm Reinforcement machine learning algorithm What are the applications of machine learning ? OR List the application of machine learning. AKTU 2022-23, Marks 02 Applications of machine learning are : age recognition. . Speech recognition Medical diagnosis . Statistical arbitrage arning association What are the advantages of machine learning ? Advantages of machine learning : ‘asily identifies trends and patterns. © human intervention is needed. ~~ SQ-2P (MCA-Sem-4) 2M 3. Continuous improvement. . Handling multi-dimensional and multi-variety data. . What are the disadvantages of machine learning ? Disadvantages of machine learning : Data acquisition Time and resources Interpretation of results High error-susceptibility What is the role of machine learning in human life 2 Role of machine learning in human life : Learning Reasoning Problem solving Language understanding What are the components of machine learning system ? Components of machine learning system are : Sensing Segmentation Feature extraction Classification Post processing What are the classes of problem in machine learning ? Classes of problem in machine learning are : Classification Regression Clustering Rule extraction List out the four issues in machine learning. AKTU 2021-22, Marks 02 Issues related with machine learning are : Data quality Transparency Traceability Reproduction of results . Explain supervised learning with an example. AKTU 2022-23, Marks 02 jati ing, in which Supervised learning is also known as associative learning, In W the network is trained by providing it with input and matching output patterns. Example : Email filtering and voice recognition. cerca , re earning SQ-3 P (MCA-Sem-4) ioe fine unsupervised learning ? gale n supervised learning is also known as self-organization, in which poe output unit is trained to respond to cj an 12. 1.15. 1.16. we Ee 17. 'usters of pattern within the input. pefine well defined learning problem, is said to learn from i i mputer program is sai experience E with respect argme class of tasks T and performance measure P, it ‘ts performance at tasks in T, as measured by P, improves with experience E. what are the features of learning problems ? Features of learning problems are ; The class of tasks (T). The measure of performance to be improved (P), The source of experience (E). . Define decision tree learning. Decision tree learning is the predictive modeling approaches used in statistics, data mining and machine learning. It uses a decision tree to go from observations about an item to conclusions about the item’s target values. What is decision tree ? Adecision tree is a decision support tool that uses a tree-like model of decisions and their possible consequences, including chance event outcomes, resource costs and utility. What are the types of decision tree ? There are two types of decision tree : - Classification tree Regression tree Define classification tree and regression tree. Classification tree : A classification tree is an algorithm where the target variable is fixed. This algorithm is used to identify the class within which a target variable would fall. ‘“Bression tree : A regression tree is an algorithm where the vareet variable is not fixed and this algorithm is used to predict its value, Name different decision tree algorithm. hitterent decision tree algorithms are : C4.5 CART ~~ ESS’ " SQ4P (MCASem.4) 2 Marks Questions 1.19. What are the issues related with the decision tree ? OR Mllustrate the issues in decision tree learning. AKTU 202: Issues related with decision tree are : Missing data . Multi-valued attribute Continuous and integer valued input attributes Continuous-valued output attributes What are the attribute selection measures used in decision tree? Attribute selection measures used in decision tree are : Entropy . Information gain Gain ratio 121. Compare artificial intelligence and machine learning. S.No.| Artificialintelligence | Machine learning —_| 1L Artificial intelligence is a | Machine learning is a subset a technology which enables a | AI which allows a machine to machine to simulate human | automatically learn from past behavior | data without programming explicitly. The goal of Al is to make a The goal of ML is to allow smart computer system like | machines to learn from data so ; humans to solve complex that they can give accurate problems. output. 1.22. Define clustering. AKTU 2022-23, Marks 02 Ams. Clustering is the task of dividing the unlabeled data or data points into different clusters such that similar data points fall in the same cluster than those which differ from the others ©O@

You might also like