data wrangling

更新时间：试题数量：购买人数：提供作者：

有效期：个月

章节介绍：共有个章节

我的练习

我的错题
(0道)

我的收藏
(0道)

我的斩题
(0道)

我的笔记
(0道)

专项练习

顺序练习 练习进度 0 / 0

随机练习 自定义设置练习量

模拟考试 仿真模拟

题型练习 按题型分类练习

易错题 精选高频易错题

学习资料 考试学习相关信息

搜索

题库预览

Briefly describe (in one or two sentences) the key characteristics of the following types of data, and provide one example for each:

(1) Categorical

(2) Numerical

(3) Ordinal

Clearly show answers for (1) to (3).

List two evaluation measures that can be problematic when used in the context of record linkage and in your own words describe why that is the case using one or two sentences for each.

Assume you have two databases and they both contain 10% errors in the attribute weekday of birth, where this attribute can contain the seven values "Monday" to "Sunday". Also assume each of the seven possible values ("Monday" to "Sunday") occurs with equal frequency in the records in both databases.

Assume you have two databases, and both contain 10% errors in the attribute age (85–90), where this attribute can take one of the six possible values: 85, 86, 87, 88, 89, or 90.

Also, assume that these six possible values (85–90) are evenly distributed in the records in both databases.

Based on the weight calculations used in probabilistic record linkage, calculate:

Agreement weight: The weight when the attribute values match between two records.

Disagreement weight: The weight when the attribute values do not match between two records.

Assume in a record linkage process the blocking method used resulted in a reduction ratio of RR=0.98, a pairs completeness of PC=0.94, and a pairs quality of PQ=0.88.

What is the maximum possible recall that this record linkage process can obtain based on these blocking results, independent of the comparison and classifications methods used?

You must round your answer to two decimal places (for example 0.42)

Assume a record linkage process has obtained a precision result of P = 0.80.

Assuming this process has correctly classified 2,380 record pairs as matches (true matches), how many record pairs did this process

incorrectly classify as matches (false matches)?

（含图）

Which of the following are data wrangling activities?

Select one or more:

a. Data collection

b. Data integration

c. Data transformation

d. Data analytics

e. Data evaluation

f. Data exploration

g. Data correction

h. Data parsing

i. Data transmission

j. Data mining

k. Data profiling

l. Data linkage

m. Data clustering

n. Data privacy

o. Data provenance

Assume a data stream with the following sequence of values (from oldest, value 6, to newest, value 14):

[6, 1, 11, 3, 9, 7, 15, 21, 8, 2, 7, 12, 18, 1, 5, 3, 15, 11, 22, 14]

Assume a window length of 10 values that can be kept in memory.

Now assume the next data value that arrives in this stream is 16.

Based on the given window size, calculate the median value of the data stream using the sliding window approach.

If required round your answer to one decimal place (like 0.7)

Imagine a data set with 9 attributes.

How many possible sub-sets of attributes containing at least three attributes can you generate using a feature selection method?

For the given list of values, L = [46, 4, 22, 31, 45, 21, 0, 47, 2, 754, 566, 7], apply robust normalisation and calculate the normalised value of the largest value in L.

Round your answer to the nearest integer value (like 42).

(For MAD calculation please use the scaling factor as 1.0)

3. What is Garbage-in garbage-out principle? Explain in one or two sentences.

Apply equal-depth binning with 3 bins for the following list of values and smooth by bin mean. Write the final list of binned values.Values:

[2, 45, 12, 32, 38, 9, 10, 57, 24, 4, 60, 51]

Apply min-max normalisation to the following list of numbers and provide the normalised list of values (round your answers to two decimal places, e.g., 27.36).

L = [9, 54, 23, 2, 33, 44, 16, 41]

.If a matcher of a schema matching application is based on metadata-based structures, what inputs/aspects can we use for this matcher?

Linking records in a single dataset is called?

Generate Soundex codes for the following two names:Graham;David

Assume you are deduplicating a database that contains 4000 records. If you are doing a naive pairwise comparison, what is the total number of record pairs that you need to compare?

2. Given the name values 'olivia' and 'olive', calculate the Jaccard similarity between those two values considering bigrams (q=2).

What is the Levenshtein edit distance between the two name values francis and frank?

(Assume a cost of 1 for all operations - deletion, insertion, and substitution.)

1 2 3 4 5 6 7