I have a number of tables that looks as follows:
time | node | left |LP iter|LP it/n|mem/heur|mdpt |vars |cons |rows |cuts |sepa|confs|strbr| dualbound | primalbound | gap | compl.
0.0s| 1 | 0 | 100 | - | 1046k | 0 | 100 | 102 | 100 | 0 | 0 | 0 | 0 | -- | 9.999990e+05*| Inf | unknown
* 0.3s| 1 | 0 | 100 | - | LP | 0 | 200 | 102 | 100 | 0 | 0 | 0 | 0 | -- | 5.587300e+04 | Inf | unknown
12.0s| 1 | 0 | 239 | - | 1781k | 0 | 239 | 102 | 100 | 0 | 0 | 0 | 0 | 5.577800e+04 | 5.587300e+04 | 0.17%| unknown
12.1s| 1 | 0 | 287 | - | 2595k | 0 | 239 | 102 | 935 | 835 | 1 | 0 | 0 | 5.577800e+04 | 5.587300e+04 | 0.17%| unknown
66.8s| 1 | 0 | 422 | - | 3061k | 0 | 336 | 102 | 935 | 835 | 1 | 0 | 0 | 5.577800e+04 | 5.587300e+04 | 0.17%| unknown
89.4s| 1 | 0 | 481 | - | 3218k | 0 | 361 | 102 | 935 | 835 | 1 | 0 | 0 | 5.580100e+04 | 5.587300e+04 | 0.13%| unknown
89.5s| 1 | 0 | 579 | - | 3513k | 0 | 361 | 102 |1335 |1235 | 2 | 0 | 0 | 5.580100e+04 | 5.587300e+04 | 0.13%| unknown
100s| 1 | 0 | 715 | - | 3837k | 0 | 403 | 102 |1335 |1235 | 2 | 0 | 0 | 5.583250e+04 | 5.587300e+04 | 0.07%| unknown
I'm interested in recording the first numeric value in the gap column (second last column of the table). The gap column could either have Inf or x.xx% values in it. If all the values in the gap column are Inf, then I would simply record Inf, otherwise, I would like to record the first numeric value. For e.g. in the above table, the value that I would like to record is 0.17. I tried many different ways but couldn't achieve any success. It would be really great if someone could provide some guidance as to how to achieve the above-mentioned objective. Thanks !
You may use this awk solution:
awk -F '[[:blank:]]*\\|[[:blank:]]*' '
NR > 1 && (!v || v == "Inf") {
v = ($(NF-1) == "Inf" ? $(NF-1) : $(NF-1)+0)
}
END {
print v
}' file
0.17
I have the following panel data set with very large N (500,000) and small T (15 years). My dependent variable is Project1 or project 2. I want to estimate the likelihood of Project dependent on treated with year and village fixed effects. For the continuous dependent variable, I was using reghdfe.
The dependent variable is simply that when a village gets the project the dummy is equal to 1 and remains 1 for the subsequent years.
I am aware that I cannot use "probit" command in STATA as I have a panel. Can you suggest which model should I use?
| village | population | year | project_1 | project_2 | treated |
|---------|------------|------|-----------|-----------|-----------|
| A | 100 | 2001 | 0 | 0 | 0 |
| A | 100 | 2002 | 1 | 0 | 0 |
| A | 100 | 2003 | 1 | 0 | 1 |
| A | 100 | 2004 | 1 | 0 | 1 |
| A | 100 | 2005 | 1 | 0 | 1 |
| B | 200 | 2001 | 0 | 0 | 0 |
| B | 200 | 2002 | 0 | 0 | 1 |
| B | 200 | 2003 | 0 | 1 | 1 |
| B | 200 | 2004 | 0 | 1 | 1 |
| B | 200 | 2005 | 0 | 1 | 1 |
| C | 150 | 2001 | 0 | 0 | 0 |
| C | 150 | 2002 | 0 | 0 | 0 |
| C | 150 | 2003 | 0 | 0 | 0 |
| C | 150 | 2004 | 1 | 0 | 0 |
| C | 150 | 2005 | 1 | 0 | 1 |
| D | 175 | 2001 | 0 | 0 | 0 |
| D | 175 | 2002 | 0 | 0 | 0 |
| D | 175 | 2003 | 0 | 0 | 0 |
| D | 175 | 2004 | 0 | 0 | 1 |
| D | 175 | 2005 | 0 | 0 | 1 |
Your question has two parts. Which model of Logit and Probit is more appropriate for you, and how to implement the appropriate model in Stata. As #NickCox mentioned, the former is most appropriate for Cross Validated, and has received robust discussion there: Difference between logit and probit models
.
I have some hardware IPs that I need to synthesize. And the IP contains several generic parameters I can play with. Each combination of parameters gives me a different utilization report after synthesis and implementation.
So for example for two different configurations Design_1 and Design_2, I get the following in Vivado 2018.1. The 3rd line is the ratio of the values of Design_2 devided by values of Design_1.
So as you can see in this simple example, Design_2 has less Slice LUTs but slightly more F7 Muxes.
My question is how to conclude about the cost of each one? Should I privilege Slice LUTs or Registers ...etc?
+----------+-------------------+-----------------+------------------+----------+-------------------+-------------------+---------------+---------------------+----------------+------+------------+--------------+-------------+------------+----------+---------+------------+---------+---------------------------+-------------------------+-----------------------------+--------+--------+----------+---------+------------+-----------+---------+--------+---------+---------+-----------+----------+-----------+-------------+---------+----------+-----------+---------+
| Name | Slice LUTs | Slice Registers | F7 Muxes | F8 Muxes | Slice | LUT as Logic | LUT as Memory | LUT Flip Flop Pairs | Block RAM Tile | DSPs | Bonded IOB | Bonded IPADs | PHY_CONTROL | PHASER_REF | OUT_FIFO | IN_FIFO | IDELAYCTRL | IBUFDS | PHASER_OUT/PHASER_OUT_PHY | PHASER_IN/PHASER_IN_PHY | IDELAYE2/IDELAYE2_FINEDELAY | ILOGIC | OLOGIC | BUFGCTRL | BUFIO | MMCME2_ADV | PLLE2_ADV | BUFMRCE | BUFHCE | BUFR | BSCANE2 | CAPTUREE2 | DNA_PORT | EFUSE_USR | FRAME_ECCE2 | ICAPE2 | PCIE_2_1 | STARTUPE2 | XADC |
+----------+-------------------+-----------------+------------------+----------+-------------------+-------------------+---------------+---------------------+----------------+------+------------+--------------+-------------+------------+----------+---------+------------+---------+---------------------------+-------------------------+-----------------------------+--------+--------+----------+---------+------------+-----------+---------+--------+---------+---------+-----------+----------+-----------+-------------+---------+----------+-----------+---------+
| Design_1 | 34124 | 16913 | 1453 | 91 | 10272 | 31538 | 2586 | 9020 | 37 | 11 | 125 | 0 | 1 | 1 | 4 | 2 | 1 | 0 | 4 | 2 | 16 | 16 | 46 | 10 | 0 | 2 | 2 | 0 | 2 | 0 | 4 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Design_2 | 34097 | 16913 | 1550 | 91 | 10189 | 31511 | 2586 | 9021 | 37 | 11 | 125 | 0 | 1 | 1 | 4 | 2 | 1 | 0 | 4 | 2 | 16 | 16 | 46 | 10 | 0 | 2 | 2 | 0 | 2 | 0 | 4 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| -------- | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| (2)/(1) | 0.999208768022506 | 1 | 1.06675843083276 | 1 | 0.991919781931464 | 0.999143889910584 | 1 | 1.00011086474501 | 1 | 1 | 1 | #DIV/0! | 1 | 1 | 1 | 1 | 1 | #DIV/0! | 1 | 1 | 1 | 1 | 1 | 1 | #DIV/0! | 1 | 1 | #DIV/0! | 1 | #DIV/0! | 1 | #DIV/0! | #DIV/0! | #DIV/0! | #DIV/0! | #DIV/0! | #DIV/0! | #DIV/0! | #DIV/0! |
+----------+-------------------+-----------------+------------------+----------+-------------------+-------------------+---------------+---------------------+----------------+------+------------+--------------+-------------+------------+----------+---------+------------+---------+---------------------------+-------------------------+-----------------------------+--------+--------+----------+---------+------------+-----------+---------+--------+---------+---------+-----------+----------+-----------+-------------+---------+----------+-----------+---------+
It's depending on your needs, LUTs and F7 Muxes are differents physical cells in your FPGA. So even if you don't use its, its will be there.
If you have one ressource more critical than the other, you should try to minimize the utilisation of the critical ressource to simplify the place and route.
If you have nothing critical, I think the better is to use F7 Muxes first because Slice LUTs are more flexible for the rest of your design.
Given two sets, e.g.:
{A B C}, {1 2 3 4 5 6}
I want to generate the Cartesian product in an order that puts as much space as possible between equal elements. For example, [A1, A2, A3, A4, A5, A6, B1…] is no good because all the As are next to each other. An acceptable solution would be going "down the diagonals" and then every time it wraps offsetting by one, e.g.:
[A1, B2, C3, A4, B5, C6, A2, B3, C4, A5, B6, C1, A3…]
Expressed visually:
| | A | B | C | A | B | C | A | B | C | A | B | C | A | B | C | A | B | C |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 1 | | | | | | | | | | | | | | | | | |
| 2 | | 2 | | | | | | | | | | | | | | | | |
| 3 | | | 3 | | | | | | | | | | | | | | | |
| 4 | | | | 4 | | | | | | | | | | | | | | |
| 5 | | | | | 5 | | | | | | | | | | | | | |
| 6 | | | | | | 6 | | | | | | | | | | | | |
| 1 | | | | | | | | | | | | | | | | | | |
| 2 | | | | | | | 7 | | | | | | | | | | | |
| 3 | | | | | | | | 8 | | | | | | | | | | |
| 4 | | | | | | | | | 9 | | | | | | | | | |
| 5 | | | | | | | | | | 10| | | | | | | | |
| 6 | | | | | | | | | | | 11| | | | | | | |
| 1 | | | | | | | | | | | | 12| | | | | | |
| 2 | | | | | | | | | | | | | | | | | | |
| 3 | | | | | | | | | | | | | 13| | | | | |
| 4 | | | | | | | | | | | | | | 14| | | | |
| 5 | | | | | | | | | | | | | | | 15| | | |
| 6 | | | | | | | | | | | | | | | | 16| | |
| 1 | | | | | | | | | | | | | | | | | 17| |
| 2 | | | | | | | | | | | | | | | | | | 18|
or, equivalently but without repeating the rows/columns:
| | A | B | C |
|---|----|----|----|
| 1 | 1 | 17 | 15 |
| 2 | 4 | 2 | 18 |
| 3 | 7 | 5 | 3 |
| 4 | 10 | 8 | 6 |
| 5 | 13 | 11 | 9 |
| 6 | 16 | 14 | 12 |
I imagine there are other solutions too, but that's the one I found easiest to think about. But I've been banging my head against the wall trying to figure out how to express it generically—it's a convenient thing that the cardinality of the two sets are multiples of each other, but I want the algorithm to do The Right Thing for sets of, say, size 5 and 7. Or size 12 and 69 (that's a real example!).
Are there any established algorithms for this? I keep getting distracted thinking of how rational numbers are mapped onto the set of natural numbers (to prove that they're countable), but the path it takes through ℕ×ℕ doesn't work for this case.
It so happens the application is being written in Ruby, but I don't care about the language. Pseudocode, Ruby, Python, Java, Clojure, Javascript, CL, a paragraph in English—choose your favorite.
Proof-of-concept solution in Python (soon to be ported to Ruby and hooked up with Rails):
import sys
letters = sys.argv[1]
MAX_NUM = 6
letter_pos = 0
for i in xrange(MAX_NUM):
for j in xrange(len(letters)):
num = ((i + j) % MAX_NUM) + 1
symbol = letters[letter_pos % len(letters)]
print "[%s %s]"%(symbol, num)
letter_pos += 1
String letters = "ABC";
int MAX_NUM = 6;
int letterPos = 0;
for (int i=0; i < MAX_NUM; ++i) {
for (int j=0; j < MAX_NUM; ++j) {
int num = ((i + j) % MAX_NUM) + 1;
char symbol = letters.charAt(letterPos % letters.length);
String output = symbol + "" + num;
++letterPos;
}
}
What about using something fractal/recursive? This implementation divides a rectangular range into four quadrants then yields points from each quadrant. This means that neighboring points in the sequence differ at least by quadrant.
#python3
import sys
import itertools
def interleave(*iters):
for elements in itertools.zip_longest(*iters):
for element in elements:
if element != None:
yield element
def scramblerange(begin, end):
width = end - begin
if width == 1:
yield begin
else:
first = scramblerange(begin, int(begin + width/2))
second = scramblerange(int(begin + width/2), end)
yield from interleave(first, second)
def scramblerectrange(top=0, left=0, bottom=1, right=1, width=None, height=None):
if width != None and height != None:
yield from scramblerectrange(bottom=height, right=width)
raise StopIteration
if right - left == 1:
if bottom - top == 1:
yield (left, top)
else:
for y in scramblerange(top, bottom):
yield (left, y)
else:
if bottom - top == 1:
for x in scramblerange(left, right):
yield (x, top)
else:
halfx = int(left + (right - left)/2)
halfy = int(top + (bottom - top)/2)
quadrants = [
scramblerectrange(top=top, left=left, bottom=halfy, right=halfx),
reversed(list(scramblerectrange(top=top, left=halfx, bottom=halfy, right=right))),
scramblerectrange(top=halfy, left=left, bottom=bottom, right=halfx),
reversed(list(scramblerectrange(top=halfy, left=halfx, bottom=bottom, right=right)))
]
yield from interleave(*quadrants)
if __name__ == '__main__':
letters = 'abcdefghijklmnopqrstuvwxyz'
output = []
indices = dict()
for i, pt in enumerate(scramblerectrange(width=11, height=5)):
indices[pt] = i
x, y = pt
output.append(letters[x] + str(y))
table = [[indices[x,y] for x in range(11)] for y in range(5)]
print(', '.join(output))
print()
pad = lambda i: ' ' * (2 - len(str(i))) + str(i)
header = ' |' + ' '.join(map(pad, letters[:11]))
print(header)
print('-' * len(header))
for y, row in enumerate(table):
print(pad(y)+'|', ' '.join(map(pad, row)))
Outputs:
a0, i1, a2, i3, e0, h1, e2, g4, a1, i0, a3, k3, e1,
h0, d4, g3, b0, j1, b2, i4, d0, g1, d2, h4, b1, j0,
b3, k4, d1, g0, d3, f4, c0, k1, c2, i2, c1, f1, a4,
h2, k0, e4, j3, f0, b4, h3, c4, j2, e3, g2, c3, j4,
f3, k2, f2
| a b c d e f g h i j k
-----------------------------------
0| 0 16 32 20 4 43 29 13 9 25 40
1| 8 24 36 28 12 37 21 5 1 17 33
2| 2 18 34 22 6 54 49 39 35 47 53
3| 10 26 50 30 48 52 15 45 3 42 11
4| 38 44 46 14 41 31 7 23 19 51 27
If your sets X and Y are sizes m and n, and Xi is the index of the element from X that's in the ith pair in your Cartesian product (and similar for Y), then
Xi = i mod n;
Yi = (i mod n + i div n) mod m;
You could get your diagonals a little more spread out by filling out your matrix like this:
for (int i = 0; i < m*n; i++) {
int xi = i % n;
int yi = i % m;
while (matrix[yi][xi] != 0) {
yi = (yi+1) % m;
}
matrix[yi][xi] = i+1;
}
I am very new to statsample and having some basic questions. With this sample data:
[[1, 2, 3, 3],[2, 3, 3, 5],[4, 1, 3, 4]]
I create a 4x4 statsample dataaset called ds and get the following output for each call:
puts ds.summary
gets
= Dataset 1
Cases: 3
Element:[actuals]
== Vector 3
n :3
n valid:3
factors:3
mode: 3
Distribution
+---+---+---------+
| 3 | 3 | 100.00% |
+---+---+---------+
Element:[mids]
== Vector 2
n :3
n valid:3
factors:1,2,3
mode: 2
Distribution
+---+---+--------+
| 1 | 1 | 33.33% |
| 2 | 1 | 33.33% |
| 3 | 1 | 33.33% |
+---+---+--------+
Element:[predicteds]
== Vector 4
n :3
n valid:3
factors:3,4,5
mode: 3
Distribution
+---+---+--------+
| 3 | 1 | 33.33% |
| 4 | 1 | 33.33% |
| 5 | 1 | 33.33% |
+---+---+--------+
Element:[prediction_error]
== Vector 5
n :3
n valid:3
factors:0,1,2
mode: 0
Distribution
+---+---+--------+
| 0 | 1 | 33.33% |
| 1 | 1 | 33.33% |
| 2 | 1 | 33.33% |
+---+---+--------+
Element:[uids]
== Vector 1
n :3
n valid:3
factors:1,2,4
mode: 1
Distribution
+---+---+--------+
| 1 | 1 | 33.33% |
| 2 | 1 | 33.33% |
| 4 | 1 | 33.33% |
+---+---+--------+
Which seems reasonable but then:
cm = ds.correlation_matrix
puts cm.summary
gets this, which is confusing:
Correlation Matrix
+------------------+---------+-------+------------+------------------+-------+
| | actuals | mids | predicteds | prediction_error | uids |
+------------------+---------+-------+------------+------------------+-------+
| actuals | 1.000 | -- | -- | -- | -- |
| mids | -- | 1.000 | -- | -- | -- |
| predicteds | -- | -- | 1.000 | -- | -- |
| prediction_error | -- | -- | -- | 1.000 | -- |
| uids | -- | -- | -- | -- | 1.000 |
+------------------+---------+-------+------------+------------------+-------+
You created a dataset with nominal vectors, not scalar ones. So, correlations between not numeric vectors is always 0.