What does a Product Manager Do?

"What does a Product Manager Do?"

My own understanding is that a Product Manager's job includes:
0) Product's Core Value and Competition
1) Product Architecture
2) Product Life Cycle or Time Table (and management)
3) Communication
4) Also, in my position, I need to code all algorithms and data I designed

Then, I searched the website on the topic, and get something:
From: http://www.smashingmagazine.com/2014/09/17/why-companies-need-full-time-product-managers/

With that as the backdrop, my definition of the role of a product manager would be to achieve business success by meeting user needs through the continual planning and execution of digital product solutions.

...

Marty Cagan has a great list of common tasks that product managers are responsible for in his ebook Behind Every Great Product (PDF). The tasks include:

1. identifying and assessing the validity and feasibility of product opportunities,
2. making sure the right product is delivered at the right time,
3. developing a product strategy and road map for development,
4. leading the team in executing the product’s road map,
5. evangelizing the product internally to the executive team and colleagues,
6. representing customers through the product development process.

Machine Learning Course 8 Study Log (Neural Network II)

(1)

(2)

(3)

(4) The Back proragation Algorithm

(5)

(6)

(7) Verify Gradient Calculation

(8)

(9)

(10)

(11) Problem of 0 as starting points

(12)

(13) All-In-One

(14) All-In-One

Machine Learning Course 5 Study Log (Matlab)

1:
A = 1;
A == 1;
A = B;
1 && 0
xor(0, 1)
b = 'hi';
c = ( 3 > 1);
disp(sprinf('%0.2f',a));
format short/long
A = [2 3; 4 5];
A = [0:0.1:1];
A = ones(2,3);
A = zeros(2,3);
A = rand(2,3); %uniform
A = randn(2,3); %normally distributed random numbers
A = eye(2,3); %identity matrix

2:
size()
length()
pwd
ls
load or importdata
who or whos
clear
A = B(1:2)
A = B(1:2,:)
A(:,2)
A(:,[1 2])
A(:,[1:1:3])
A(1:2,:) = [1; 2]
C = [A; B]
C = [A B] or [A, B]
save 'text.txt' A - ascii
save 'text.mat' A %binary format

3:
log(A) or abs(A)
A + 2
A * 2 or A .* 2
-A
A * B
A .* B
(A)'
pinv(A)
A < 3
find(A < 3)
sum(sum(A))
prod(A)
floor , ceil
max(rand(3), rand(3))
max(A, [], 1)
A(:)
flipud

4:
plot()
plot(x,y,'r')
hold on;
xlabel
ylabel
legend
title
print -dpng 'test.png'
figure(1) %define which figure to use
subplot(1, 2, 1) % cut the plot into 1 * 2 and then use 1
imagesc, colorbar, colormap gray

5:
for i in 1:0.1:10
  XXX
end

break;
continue;

while i < 10
  XXX
end

if
elseif
elseif
else
end

function [y1, y2] = a_function_name(x)
   y1 = x^2
   y2 = x^3


search path; addpath ('')

6:

7.

Machine Learning Course 8 Study Log (Neural Network)

(1)

(2)

(3)

(4)

The Idea under Neural Network is to Learn Required Composed Features As it Wants

(5)

(6)

(7)

(8)

Machine Learning Course 7 Study Log (Regularization)

(1)

(2)

(3)

(4)

Machine Learning Course 6 Study Log (Logistic Regression)

(1) Why linear regression is not good?

(2)

(3)

(4)

(5)

(6)

(7)

Machine Learning Course 4 Study Log (Linear Regression with Multiple Variables)

(1)

(2)

(3)

(4)

(5)

Machine Learning Course 3 Study Log (Linear Algebra Review)

Matrix Inverse:
A * A-1 = I (identity matrix)

Matrix Transpose:
Aij = Aji

Machine Learning Course 2 Study Log (Linear Regression with One Variable)

(1)

（1.5）

(2)

(3)

(4)

Machine Learning Course 1 Study Log (Concept)

Supervised Learning

In supervised learning, we are given a data set and already know what our correct output should look like, having the idea that there is a relationship between the input and the output.

Supervised learning problems are categorized into "regression" and "classification" problems. In aregression problem, we are trying to predict results within a continuous output, meaning that we are trying to map input variables to some continuous function. In a classification problem, we are instead trying to predict results in a discrete output. In other words, we are trying to map input variables into discretecategories.

Example:

Given data about the size of houses on the real estate market, try to predict their price. Price as a function of size is a continuous output, so this is a regression problem.

We could turn this example into a classification problem by instead making our output about whether the house "sells for more or less than the asking price." Here we are classifying the houses based on price into two discrete categories.

Unsupervised Learning

Unsupervised learning, on the other hand, allows us to approach problems with little or no idea what our results should look like. We can derive structure from data where we don't necessarily know the effect of the variables.

We can derive this structure by clustering the data based on relationships among the variables in the data.

With unsupervised learning there is no feedback based on the prediction results, i.e., there is no teacher to correct you. It’s not just about clustering. For example, associative memory is unsupervised learning.