Please enable JavaScript.

Coggle requires JavaScript to display documents.

Computer Science (Algorithms (Operations to construct: (Sequential:…

- - - - Unambiguous: understood and carried out directly without having further simplification or explanation = PRIMITIVE OPERATION, tedious: higher level primitives
      - e. Computable: doable by the computing agent
  - - - Time
      - Space
- - - - related fields: machine learning; visualisation, databases, statistics
      - Metodology: 1. data >select 2, target data >pre process (remove, outliers) 3. Preprocessed data >transfer, (reduce dimensionality) 4. transformed data > data mine (find patterns) 5. patterns/models > interpret 5. knowledge
    - - Classification i.e. fraud detection find classifiers for class
      - Regression: predict value of variable based on values of other variables
      - Clustering: Given set of points, ,easier points more similar >in one cluster. i.e. market segmentation: collect attributes, find clusters, measure quality by observing patterns, personalise campaign + Recommendation system
      - Association rule task: Produce dependency rules to predict occurances. >plan distribution (supermarket co-occurences: when being milk, also coke (no causality!)
- - - - more productive user friendly, low level programming language, each symbol in assembly language is translated to binary language instructions
      - program in assembly: source program ; translation to machine language: object program (assembler)
      - Source program >assembler>object program >loader > object pr. loaded into memory >hardware>results
      - Instruction format:
        
        Label: op code- address --comments
        
        directives: .begin .end
        
        labeling instruction or data: namelable: load x
        
        generate data: x = .data +5 ( label equiv. to address)
      - .begin .."instructions" .. halt .."terminate execution, data generation... end
- - - - holds operands or stores results of computation
      - connected by (bus). to left and right operands
  - - - Machine instruction set depends on numbers of possible op codes. RISC: Reduced instruction set computers (30-50) & CISC Complex (300-500)
      - Codes:
        
        Data moevement: -load: content of address into R, Store: copy content of R into
        
        Arithmetic: Add content of R into call at address C and leave sum in R
        
        Comparisons: load 2nd to R , compare to first: memory cell vs. data register R. result deposited in condition code register CCR
        
        Jump: Check condition code and transfer control to instruction at location c in the code set in a certain way
  - - - DASDs: direct access storage devices: every unit of info has a unique address but tie needed to access depends on location. i.e. end approx. constant time
      - SADs sequential storage devices when a memory accessed, all previous cells accessed too. Time depends on number of cells. i.e. tape devices. Useful when you have to process all info have. i.e. backup info
      - Human readable: Keyboard, screen, printers; Machine readable: disks, tapes .. non volatile!
- - - - Pre- pruning: (easy, less efficient): stop growing tree earlier
      - Post-proning: Allow overfitting, analyse loss when pruning a breach, then prune
      - In this case it is possible that for coincidental regularities to occur that are unrelated to the actual borders.
- - - - Fortan (Formula Translation): no while, use Go to for loops> difficult spaghetti code, mixed language program still on use, (was first) support numerical computations
      - Cobol ( Common business oriented language) : managing inventories and protocols. update master files with changes from transaction files >summary reports. File input (not keyboard). Y2k problem (formatting storages for dates beginning 2020)
      - Pascal : enforces programming techniques, smilar to pseudocode
      - C: system programming but now general purpose language: use low level info (close to assembly)
    - - based on combination of transformations upon items, every task in terms of functions
      - LIS
        
        processes lists of things incl. arithmetic, comp, logical, bitwise
        
        fundamental functions: CAR (content address register; non empty, returns first element) ; COR(content degree register) , takes non empty list and return without first element) ; NULL? returns sure is list is empty, false otherwise: ADDER: adds another layer of abstraction, pure math, clarity go thought
    - - logical deductions from known facts; not specified how the task is done. DECLARATIVE language. express property of objects, or relationship long several.
      - Facts asserted to be true, to infer on other facts
      - Question? > begins with facts and apply logic deduction to answer query
      - write expert systems; also DECLARATIVE since instead of issuing commands, decrlarations of true facts
      - Query >inference engine>response ; rules + facts = knowledgebvase >>inference engine
    - - multicore computing c multiple processors on single circuit
      - cluster computing (independent system laptops, interconnected by network (internet)
      - : parallel execution of subtasks of problem on different processors ; provide computational speed required to solve large scale problems
      - "grand challenge " computing problems i.e whether prediction
      - SIMD (single instruction, multiple data strea,: each .. carries out instruction in its loc al data, each pu in a different part of array.
      - MIMD : multiple instruction : numerous interconnected processors execute own program in own data. i.e each processor looks for target value in diff.part of data set. when one finds it, send "interrupt" signal the rest to halt